The Eastern Cooperative Oncology Group (ECOG) conducted a prospective phase III study in patients with relapsed/refractory acute myeloid leukemia (AML) to evaluate whether administration of repeated courses of low-dose cytarabine (LDAC) maintenance therapy after induction of complete remission in advanced AML would improve disease-free and overall survival. Patients with AML in second/later relapse or refractory disease were first treated with a combination of high-dose cytarabine and amsacrine. Those who achieved complete remission were then randomized to observation or to receive LDAC, 10 mg/m2subcutaneously twice a day ×2 21 days every 2 months until relapse occurred. Of 86 patients eligible for randomization, 41 patients were assigned to receive LDAC and 45 patients to observation. The median disease-free survival was 7.4 months for patients assigned to LDAC compared to 3.3 months for patients receiving no additional therapy, P = 0.084. The median survival from randomization was 10.9 months and 7.0 months for patients receiving LDAC maintenance chemotherapy and observation, respectively (P = 0.615). The data from this study suggest that LDAC maintenance therapy given to patients with advanced AML who achieve complete remission can increase disease-free survival compared to observation, but does not improve overall survival. Nevertheless, because of the ineffectiveness and toxicity of intensive post-remission chemotherapy in this circumstance, LDAC maintenance therapy, a tolerable outpatient regimen, offers the potential for improved quality of life.
In adults with de novo, newly diagnosed acute myeloid leukemia (AML) induction chemotherapy results in a complete remission (CR) rate of approximately 65% overall, with the likelihood of CR varying inversely with patient age.1 Among those patients who achieve a CR, approximately 35% experience a durable remission. The great majority of patients therefore relapse. The likelihood of such patients achieving another CR is approximately 40%.2 Even when these patients do achieve a CR, the median duration of remission is only approximately half as long as the preceding remission, and long-term survival and cure is only obtainable after autologous or allogeneic bone marrow transplantation.2 Whether any chemotherapy less intensive than that employed in bone marrow transplantation can prolong later remissions in advanced AML is unclear. The fact that many patients with advanced disease are elderly and/or variably compromised by complications from previous treatment raised the question of whether low-dose maintenance chemotherapy would be tolerable and of benefit. Based on data demonstrating some efficacy for low-dose cytarabine (LDAC) in the therapy of AML34567 and myelodysplasia,89 a LDAC maintenance program was used in a randomized clinical trial (EST5483) of the Eastern Cooperative Oncology Group (ECOG). In order to standardize the chemotherapy regimen used to induce CR, all patients received a course of high-dose cytarabine and m-AMSA (4′(9-acridinylamine) methane sulfon-m-anisidide). Those patients who achieved remission were then randomized to receive repeated courses of low-dose cytarabine (maintenance therapy) or no further therapy (observation). The primary endpoints of this study were disease-free and overall survival.
Patients and methods
Adult patients with relapsed or refractory AML of French–American–British (FAB) types M1–M710 were enrolled on ECOG study EST 5483. Eligible patients were required to satisfy all of the following criteria: age 18–75 years old; at least one prior leukemia induction therapy; no prior history of myelodysplasia; ⩾5 weeks since the last chemotherapy; afebrile for >72 h off antibiotics; and adequate hepatic (bilirubin ⩽2.0 mg/dl, serum transaminases and alkaline phosphatase <3× the upper limit of the normal range) and renal (serum creatinine ⩽2.0 mg/dl) function. Patients were excluded from eligibility for the following reasons: prior therapy with either high-dose cytarabine (>1.5 g/m2/dose) or m-AMSA; prior malignancy for which chemotherapy or radiation therapy was administered; uncontrolled angina, congestive heart failure or cardiac disease requiring digoxin, diuretics (other than for hypertension) or antiarrhythmic agents; and prior cranial or axial radiotherapy. Patients with unstable cardiac status, myocardial infarction within the past 3 months, or in blast crisis of chronic myelogenous leukemia were ineligible.
The induction chemotherapy regimen for all patients consisted of cytarabine 3000 mg/m2 in 250 ml D5W administered over 1 h intravenously every 12 h for 6 consecutive days, followed by m-AMSA 100 mg/m2 in 250 ml D5W given as a 1-h intravenous infusion daily ×3 days (on days 7, 8 and 9).
To qualify for the randomization step of this trial, patients had to demonstrate an ECOG performance status of 0–2, continuing documentation of CR (neutrophils ⩾1500/μl and platelets ⩾100 000/μl) at the time of randomization, bilirubin ⩽2.0 mg/dl, and alkaline phosphatase and transaminase ⩽3× normal. Patients with persisting infection were ineligible.
For those patients randomized to the low-dose (LDAC) maintenance therapy, treatment began 1–3 weeks after documented CR. Investigators could delay, however, longer than 3 weeks to register patients for randomization if necessary to allow time for the patients to recover from the toxicities of induction therapy and/or to meet the eligibility requirements for randomization. Treatment consisted of cytarabine 10 mg/m2 subcutaneously every 12 h × 21 days (self-administered at home). The treatment cycle was repeated every 8 weeks until relapse. The scheduled 21 day course of treatment could be modified as follows: if the neutrophil count was <1000/μl or platelet count <50000/μl prior to day 14, cytarabine was temporarily halted and blood counts were repeated every 2 days. When the neutrophils again exceeded the threshold values, cytarabine was resumed and continued until day 21. The days missed due to low counts were deleted from that cycle. If the low blood counts occurred on or after day 14, that cycle was terminated. The next cycle was resumed 5 weeks later. If abnormal blood counts persisted, a bone marrow examination was performed.
CR was defined as an M1 marrow status (<5% blasts, >15% erythroid elements, >25% normal granulocyte precursors in an at least moderately cellular marrow), P1 physical status (no leukemic involvement of the spleen, liver or lymph nodes) and H1 peripheral blood (hemoglobin >11 g/dl; neutrophils >1500/μl, no circulating blasts; and platelets >100 000/μl). Patients who had an M1, P1 and H1 status were still considered to be in CR even if the hemoglobin was <11 mg/dl. Relapse was defined as the appearance of circulating leukemia cells or >5% blasts in the bone marrow. Toxicity was evaluated using ECOG criteria.11
A total of 404 patients were initially registered for induction therapy between June 1984 and December 1988, when the study accrual goal was reached. Of these patients, 16 failed to provide adequate documentation of diagnosis (eligibility required submission of peripheral blood and bone marrow slides for centralized review and confirmation of diagnosis), 10 patients were ineligible based on medical guidelines for study entry, nine patients were diagnosed with acute lymphocytic leukemia, six had received therapy too recently, three patients had been previously treated with high-dose cytarabine, and one patient each was ineligible because of no prior leukemia therapy, chronic myeloid leukemia, cancellation, and registration on study after treatment had already begun. This left 356 eligible patients.
The median age of the 356 eligible patients was 47 years with a range of 18–74 years. The ratio of male:female patients was 55:45, 80% had performance status 0 or 1, and only 2% of the patients had any prior radiotherapy. Only 10% of patients had failed to achieve a CR after ⩾2 two prior induction therapy attempts, and 14% were primarily refractory to treatment. The distribution of the FAB types among the eligible patients was as follows: M1, 109 (31%); M2, 101 (28%); M3, 17 (5%); M4, 71 (20%); M5, 46 (13%); M6, 8(2%); and M7, 4(1%).
Median leukemia-free survival for the patients in CR followed with observation was estimated to be 4.5 months. The study design called for 90 complete responders to be randomized. This number provided at least 90% power to detect a 100% increase in leukemia-free survival among complete responders who received LDAC maintenance therapy, using a two-sided, α = 0.05 log-rank test. Based on the assumption that approximately 22% of the patients on the induction step were evaluable for the maintenance step, a total of 405 patients were to be registered for the induction step to secure 90 evaluable complete responders for randomization.
Complete responders were randomized according to the permuted block randomization scheme, stratified by the number of failed prior induction attempts (0 vs 1 vs 2 or more) and by the objective response to prior induction attempts (CR vs other). Survival distributions for the leukemia-free survival and overall survival were estimated according to the methods of Kaplan and Meier.12 In prognostic factor analysis, the logistic regression model13 was used to assess the association between the objective response and baseline patient characteristics, and the proportional hazards regression model14 was used to assess the association between survival time and baseline patient characteristics.
For comparison of objective response rates, a two-sided Fisher's exact test13 was used for dichotomous covariates, and a Wald test was used for continuous covariates in logistic regression models. For comparison of the survival distributions, a two-sided log-rank test of Mantel15 was used for discrete covariates, and a Wald test was used for continuous covariates in proportional hazard regression models. Both tests were stratified by the stratification factors, unless otherwise indicated.
Of the 356 eligible patients who received induction therapy with high-dose cytarabine and m-AMSA, 150 or 42% (95% confidence interval, 37–47%) achieved a CR. Using the median age of 47 years as a cutpoint, the CR rate was 50% vs 34% in younger vs older patients, respectively (P = 0.04). The CR rate for patients ⩾60 years old was 30% (25/82) vs 46% (125/274) for younger patients (P = 0.015). In this patient population with advanced disease, 100 patients (28%) died of complications of induction therapy, most frequently (in 84 patients) due to infection. Because of the high treatment-related mortality rate of this induction chemotherapy, the combination of high-dose cytarabine and amsacrine given here is not recommended for subsequent use.
Of the 150 patients in CR, a total of 86 patients was eligible and entered the randomization after CR. Sixty-four patients who achieved CR were not eligible for randomization because of the following: relapse (34) or death (five) prior to randomization; exit from the study to undergo bone marrow transplantation (14 patients); delayed documentation of CR (four patients); failure to meet medical criteria for randomization eligibility (six); and prolonged thrombocytopenia (one). The 34 relapses (22% of the CRs) occurred in patients whose physicians delayed randomization awaiting patients′ recovery from induction toxicity and/or satisfaction of the eligibility criteria for the randomization. Overall, the time from CR to randomization was a median 7.5 days (range, 0–112 days). Of the 86 eligible randomized patients, one patient refused assignment to observation and opted for LDAC therapy, whereas four patients refused LDAC and three relapsed before receiving it. All of these 86 patients are included in the statistical analysis of outcome based on the intention-to-treat. No significant differences were detected between the 45 observation patients and the 41 patients receiving LDAC with regard to the distribution of prognostic and demographic features, such as age, gender, time to CR, distribution of FAB types, number of previous CRs, performance status, and peripheral blood count or bone marrow features at study entry. The groups were quite similar as shown in Table 1.
The median follow-up of these patients was 7 years and 3 months. As shown in Figure 1, the leukemia-free survival measured from the time of complete remission was 4.1 months overall; 3.7 months in patients on observation and 7.9 months on LDAC maintenance. This doubling of median disease-free survival was associated with a P value of 0.084 (log-rank) for the observed difference. When only those patients who actually received their assigned randomization are considered for analysis, the median leukemia-free survival was 3.9 months; 3.1 months for patients on observation and 7.7 months on LDAC maintenance. The observed difference became statistically significant, P = 0.027 (log-rank). Univariate analysis revealed that longer time from the start of induction therapy to CR (median value of >36 vs ⩽36 days) and longer time from CR to randomization (median value of >7.5 vs ⩽7.5 days) were associated with decreased leukemia-free survival. To assess the joint effect of individual factors on disease-free survival, a proportional hazards model was fit according to the model building technique of stepwise selection. After adjustments for time to induction CR and time to randomization, patients receiving LDAC maintenance therapy experienced statistically significant improvement in leukemia-free survival, P = 0.043. Overall survival (measured from time of randomization) of patients under observation (median, 7.0 months) and of patients receiving LDAC maintenance (median, 10.8 months) did not, however, differ significantly (P = 0.615 by unstratified logrank and P = 0.492 by stratified log-rank), as shown in Figure 2.
Of the eligible patients on observation, three patients experienced grade 3 hepatic toxicity and one patient had grade 3 infection. Of the eligible patients on LDAC maintenance, on at least one treatment cycle hematologic toxicity was grade 3 in 14 patients and grade 4 in 14 patients. Curtailment of the number of treatment days in subsequent cycles prevented repetitive severe hematotoxicity. Compliance was excellent. Thirty-one patients were treated continuously until relapse occurred for 1–24 months (median, 3 months). Only five patients in CR discontined therapy after less than 6 months of treatment. They all had relapsed less than 1 year later. Four patients in CR terminated therapy after –7½ years. One of these patients relapsed years later. The other three remained in continuous CR 3, 4, and 6½ years later, when data collection for this analysis closed. One patient remained in CR while being treated continuously with LDAC for 9½ years. Five patients receiving LDAC had grade 3 non-hematologic toxicity, including hepatic toxicity (two), vomiting (one), infection (one) and fever (one). No grade 4 (life-threatening) non-hematologic toxicity occurred, and no patient died as a result of LDAC therapy.
Escalating the intensity of therapy after first CR in AML by high-dose conventional chemotherapy16 improves remission duration, especially when compared to no further therapy17 or low-dose maintenance therapy.1 When CR is achieved in advanced AML (eg second or later remissions, refractory disease), remission duration and survival are usually brief.2 Autologous18 or allogeneic19 bone marrow transplantation can produce cures in up to 30% of such patients. It is unknown, however, whether any other post-remission therapy, except for bone marrow transplantation, can improve disease-free or overall survival in this circumstance. The problems of low CR rates and the difficulty of maintaining remission by conventional chemotherapy in advanced AML are similar to those that confront older patients in initial remission. In first remission AML in the elderly, intensive conventional chemotherapy does not provide benefit and poses unacceptable risks of morbidity and mortality.16 Older patients are over-represented among patients presenting for chemotherapy with advanced or refractory AML, because they are less likely than younger patients to be cured by initial therapy and more often unable to tolerate the rigors of bone marrow transplantation.
Since intensive therapy after remission produces substantial toxicity in patients with advanced AML, the ECOG decided to test the value of modest dosage maintenance therapy. LDAC was chosen because of its demonstrated efficacy in remission induction2021 in advanced AML34567 and myelodysplasia89 and in the chronic phase of chronic myeloid leukemia.2223 LDAC has been administered in doses ranging from 10–20 mg/m2 daily for periods of 10–20 days either subcutaneously or by continuous infusion and repeated every 3–8 weeks.2021 The most common dose and route of administration has been 10 mg/m2 every 12 h subcutaneously. Even at such minimal doses, hematotoxicity with pancytopenia and neutropenic fever resulted in mortality rates of approximately 15% when administered to patients in relapse as induction therapy.2021 In contrast, toxicity has been modest when the same doses and duration of therapy was employed in the setting of maintenance therapy after CR.3672425
The broad applicability of this study's findings is limited by the fact that only 57% of the patients achieving CR reached the randomization step. The high-risk patients relapsed before randomization. In a subset of patients studied here, maintenance therapy did not cause any lethal toxicity, and the median duration of remission associated with LDAC administered every 8 weeks until relapse was 7.4 months, more than twice as long as the observation only group. This is consistent with other studies of low-dose cytarabine therapy after CR in patients with advanced and unfavorable (with medical problems, secondary leukemia, and/or age <60 years) AML who were treated with variable course lengths every 4–6 weeks until relapse; the median duration of CR was 6–9 months.367 Archimbaud et al24 studied patients with unfavorable AML (elderly patients and secondary AML) who had achieved their initial CR with standard induction therapy. These patients received four courses (every 6 weeks) of LDAC. The median disease-free survival was 10 months. Another trial in a similar group of patients involved readministration of one additional course of induction therapy after CR.25 Patients were then randomized to receive eight courses of LDAC (10 mg/m2 subcutaneously every 12 h × 12 days every 6 weeks) or to be observed without chemotherapy. The CR duration for LDAC (median, 12 months, 20% at 3 years) was significantly better (P = 0.006) than for observation (median, 7 months, 7% at 3 years). As was true in the current study, overall survival did not differ significantly (P = 0.29) between LDAC (median 14 months, 18% at 5 years) and observation (median, 18 months, 15% at 5 years). The authors noted that the difference in disease-free survival persisted for 5 years, even though LDAC maintenance was given for only 1 year, suggesting some carry-forward of LDAC's effects on residual leukemia cells. Similar suppressive effects on leukemia regrowth were evident in Archimbaud et al's study,24 wherein there was an acceleration of the relapse rate after the 6 months of LDAC ended, at a time when the slope of the relapse curve should have been declining rather than increasing.
In conclusion, even though no improvement in survival has yet been demonstrated with LDAC maintenance, suppression of leukemia regrowth improves the disease-free survival. Because of the declining effect after discontinuation of LDAC, subsequent studies should continue LDAC maintenance therapy until relapse occurs, rather than stopping at a previously fixed interval. Patients might also benefit from reducing the opportunity for interval leukemia regrowth between courses by giving LDAC more frequently than in the previous studies, such as monthly and limiting the period of treatment to 7–10 days to allow tolerable, repeated treatment cycles. Except for bone marrow transplantation after CR induction in patients with unfavorable AML (older than 60 years, secondary disease, relapsed, poor prognosis karyotype), intensifying conventional chemotherapy (with high-dose cytarabine, for example) does not improve CR duration or survival, but it does contribute to morbidity and mortality. Given these circumstances, low-dose cytarabine (given with minimal toxicity as an outpatient regimen) has much to recommend it to improve disease-free survival, and therefore quality of life in these patients.
Cassileth PA, Lynch E, Hines JD, Oken MM, Mazza JJ, Bennett JM, McGlave PB, Edelstein M, Harrington DP, O'Connell MJ . Varying intensity of postremission therapy in acute myeloid leukemia Blood 1992 79: 1924–1930
Hiddeman W, Buchner T . Treatment strategies in acute myeloid leukemia (AML). B. Second line treatment Blut 1990 60: 163–171
Venditti A, Stasi R, Del Poeta G, Buccisano F, Aronica G, Bruno A, Pisani F, Caravita T, Masi M, Tribalto M, Simone MD, Avvisati G, Amadori S, Papa G . All-trans retinoic acid and low-dose cytarabine for the treatment of ‘poor prognosis’ acute myeloid leukemia Leukemia 1995 9: 1121–1125
Pascarella A, Marrani C, Leoni F, Ciolli S, Nozzoli C, Caporale R, Salti F, Ferrini PR . Low-dose cytosine arabinoside in patients with acute myeloid leukemia not eligible for standard chemotherapy Leuk Lymphoma 1995 18: 465–469
Slapak CA, Desforges JF, Fogaren T, Miller KB . Treatment of acute myeloid leukemia in the elderly with low-dose cytarabine, hydroxyurea, and calcitriol Am J Hematol 1992 41: 178–183
Detourmignies L, Wattel E, Lai JL, Bauters F, Fenaux P . Is there still a role for low-dose cytosine arabinoside in de novo acute myeloid leukemia in the elderly? Ann Hematol 1993 66: 235–240
Tilly H, Castaigne S, Bordessoule D, Sigaux F, Daniel M-T, Monconduit M, Degos L . Low-dose cytosine arabinoside treatment for acute nonlymphocytic leukemia in elderly patients Cancer 1985 55: 1633–1636
Griffin JD, Spriggs D, Wisch JS, Kufe DW . Treatment of preleukemic syndromes with continuous intravenous infusion of low-dose cytosine arabinoside J Clin Oncol 1985 3: 982–991
Hellstrom-Lindberg E, Robert K-H, Gahrton G, Lindberg G, Forsblom A-M, Kock Y, Ost A . Low-dose Ara-C in myelodysplastic syndromes (MDS) and acute leukemia following MDS: proposal for a predictive model Leuk Lymphoma 1994 12: 343–351
Bennett JM, Catovsky D, Daniel MT, Flandrin G, Galton DAG, Gralnick HR, Sultan C . Proposed revised criteria for the classification of acute myeloid leukemia Ann Intern Med 1985 103: 626–629
Oken MM, Davis TE, Creech RH, McFadden ET, Tormey DC, Carbone PP, Horton T . Toxicity and response criteria of the Eastern Cooperative Oncology Group Am J Clin Oncol 1982 5: 649–655
Kaplan EL, Meier P . Nonparametric estimation from incomplete observations J Am Stat Assoc 1958 53: 457–481
Cox DR, Snell EJ . Analysis of Binary Data, 2nd edn Chapman and Hall: London 1989
Cox DR, Oakes D . Analysis of Survival Data Chapman and Hall: London 1984
Mantel N . Evaluation of survival data and two new rank order statistics arising in its consideration Cancer Chemother Rep 1966 50: 163–170
Mayer RJ, Davis RB, Schiffer CA, Berg DT, Powell BL, Schulman P, Omura GA, Moore JO, McIntyre OR, Frei EIII for the Cancer and Leukemia Croup B . Intensive postremission chemotherapy in adults with acute myeloid leukemia New Engl J Med 1994 331: 896–903
Cassileth PA, Harrington DP, Hines JD, Oken MM, Mazza JJ, McGlave P, Bennett JM, O'Connell MJ . Maintenance chemotherapy prolongs remission duration in adult acute nonlymphocytic leukemia J Clin Oncol 1988 6: 583–587
Petersen FB, Lynch MHE, Clift RA, Appelbaum FR, Sanders JE, Bensinger WI, Benyunes MC, Doney K, Fefer A, Martin P, Storb R, Rowley S, Sullivan KM, Witherspoon R, Weiden P, Thomas ED, Fisher L, Hansen JA, Buckner CD . Autologous marrow transplantation for patients with acute myeloid leukemia in untreated first relapse or in second complete remission J Clin Oncol 1993 7: 1353–1360
Dinsmore R, Kirkpatrick D, Flomenberg N, Gulati S, Kapoor N, Brochstein J, Shank B, Reid A, Groshen S, O'Reilly RJ . Allogeneic bone marrow transplantation for patients with acute nonlymphocytic leukemia Blood 1984 63: 649–656
Cheson BD, Jasperse DM, Simon R, Friedman MA . A critical appraisal of low-dose cytosine arabinoside in patients with acute non-lymphocytic leukemia and myelodysplastic syndromes J Clin Oncol 1986 4: 1857–1864
Bolwell BJ, Cassileth PA, Gale RP . Low dose cytosine arabinoside in myelodysplasia and acute myelogenous leukemia: a review Leukemia 1987 1: 575–579
Robertson MJ, Tantravahi R, Griffin JD, Canellos GP, Cannistra SA . Hematologic remission and cytogenetic improvement after treatment of stable-phase chronic myelogenous leukemia with continuous infusion of low-dose cytarabine Am J Hematol 1993 43: 95–102
Guilhot F, Chastang C, Michallet M, Guerci A, Harousseau J-L, Maloisel F, Bouabdallah R, Guyotat D, Cheron N, Nicolini F, Abgrall J-F, Tanzer J for the French Chronic Myeloid Leukemia Study Group . Interferon alpha-2b combined with cytarabine versus interferon alone in chronic myelogenous leukemia New Engl J Med 1997 337: 223–239
Archimbaud E, Anglaret B, Thomas X, Jaubert J, Sebban C ., Guyotat D, Fiere D. Maintenance with low-dose cytarabine for acute myeloid leukemia in complete remission Ann Hematol 1992 65: 71–74
Lowenberg B, Suciu S, Archimbaud E, Haak H, Stryckmans P, de Cataldo R, Dekker AW, Berneman ZN, Thyss A, van der Lelie J, Sonneveld P, Visani G, Fillet G, Hayat M, Hagemeijer A, Solbu G, Zittoun R . Mitoxantrone versus daunorubicin in induction-consolidation chemotherapy – the value of low-dose cytarabine for maintenance of remission, and an assessment of prognostic factors in acute myeloid leukemia in the elderly: final report of the Leukemia Cooperative Group of the European Organization for the Research and Treatment of Cancer and the Dutch–Belgian Hemato-Oncology Cooperative Hovon Group randomized phase III study AML-9 J Clin Oncol 1998 16: 872–881
This study was conducted by the Eastern Cooperative Oncology Group (Robert L Comis, MD, Chair) and supported by Public Health Service grants CA23318, CA11083, CA13650, CA14958, CA66636, and CA21115 from the National Cancer Institute, National Institutes of Health and the Department of Health and Human Services. Its contents are solely the responsibility of the authors and do not necessarily represent the official views of the National Cancer Institute.
About this article
Cite this article
Robles, C., Kim, K., Oken, M. et al. Low-dose cytarabine maintenance therapy vs observation after remission induction in advanced acute myeloid leukemia: an Eastern Cooperative Oncology Group Trial (E5483). Leukemia 14, 1349–1353 (2000). https://doi.org/10.1038/sj.leu.2401850
- low-dose cytarabine
- acute myeloid leukemia
- maintenance therapy