Effects of post-remission chemotherapy before allo-HSCT for acute myeloid leukemia during first complete remission: a meta-analysis

Allogeneic hematopoietic stem cell transplantation (allo-HSCT) is most frequently used to treat acute myeloid leukemia (AML). Whether patients should routinely receive consolidation chemotherapy before proceeding to transplant after achieving first complete remission (CR1) has been a subject of debate. We performed a systematic review and meta-analysis of studies examining the impact of post-remission chemotherapy before allo-HSCT in patients with AML in CR1. Six studies including 1659 patients were included in the meta-analysis. The pooled hazard ratio (HR) for overall survival was 0.9 (95% confidence interval [CI] 0.77–1.05, P?=?0.182), and the pooled HR for leukemia-free survival was 0.87 (95% CI 0.75–1.0, P?=?0.07). No survival advantage was observed for post-remission chemotherapy before reduced-intensity conditioning or myeloablative conditioning (MAC) allo-HSCT for AML in CR1. The pooled relative risk for relapse incidence (RI) was 1.02 (95% CI 0.82–1.28, P?=?0.834). Post-remission chemotherapy before allo-HSCT did not significantly affect the RI in patients with AML in CR1. The analyses revealed no significant benefit of post-remission consolidation chemotherapy in patients who received allo-HSCT. We recommend proceeding to allo-HSCT as soon as CR1 is attained.     

Results of conventional-dose cytosine arabinoside and idarubicin in elderly patients with acute myeloid leukemia

Summary Conventional-dose Ara-C (200 mg/m² d 1–5) combined with idarubicin (12 mg/m² d 1–3) was employed as remission induction and consolidation therapy in 23 elderly AML patients with a median age of 66 years (range, 60–75) with AML according to the FAB criteria (M1n=3, M2n=10, M4n=6, M5n=2, M6n=2), eligible for the study. In seven patients earlier MDS had been documented by previous bone marrow aspirates. The CR rate after one induction course was 65% (15/23). Toxicity was acceptable, with four patients dying during the chemotherapy-induced hypoplasia (4/23). Although 80% of the CR patients received two additional cycles of Ara-C and idarubicin as consolidation therapy, only two patients are still in continuous complete remission more than 12 months after achieving CR. The median disease-free survival of the CR patients was 11.5 months and the median survival of the entire group was 10 months. We conclude that conventional dose Ara-C/idarubicin is an effective protocol for inducing complete remission in elderly patients with AML, but that consolidation therapy consisting of two courses of the same regimen does not produce a relevant rate of long-term disease-free survival.     

Individual patient data meta-analysis of randomized trials evaluating IL-2 monotherapy as remission maintenance therapy in acute myeloid leukemia

IL-2 is a natural, T cell-derived cytokine that stimulates the cytotoxic functions of T and natural killer cells. IL-2 monotherapy has been evaluated in several randomized clinical trials (RCTs) for remission maintenance in patients with acute myeloid leukemia (AML) in first complete remission (CR1), and none demonstrated a significant benefit of IL-2 monotherapy. The objective of this meta-analysis was to reliably determine IL-2 efficacy by combining all available individual patient data (IPD) from 5 RCTs (N = 905) and summary data from a sixth RCT (N = 550). Hazard ratios (HRs) were estimated using Cox regression models stratified by trial, with HR < 1 indicating treatment benefit. Combined IPD showed no benefit of IL-2 over no treatment in terms of leukemia-free survival (HR = 0.97; P = .74) or overall survival (HR = 1.08; P = .39). Analyses including the sixth RCT yielded qualitatively identical results (leukemia-free survival HR = 0.96, P = .52; overall survival HR = 1.06; P = .46). No significant heterogeneity was found between the trials. Prespecified subset analyses showed no interaction between the lack of IL-2 effect and any factor, including age, sex, baseline performance status, karyotype, AML subtype, and time from achievement of CR1 to initiation of maintenance therapy. We conclude that IL-2 alone is not an effective remission maintenance therapy for AML patients in CR1.     

Treatment of acute lymphoblastic leukemia in adults by the modified protocol of L-10M (Sloan-Kettering)

We have treated 20 adult patients with acute lymphoblastic leukemia (ALL) and 2 patients with lymphoblastic lymphoma with a protocol modified from L-10M of the Memorial Sloan Kettering Cancer Center. Eighteen patients (81.8%) entered complete remission (CR). Eight of them eventually relapsed (only 1 patient had a meningeal relapse) and died. Median CR duration was 19 months (median overall follow-up of 35 months and 50% remission duration was not yet determined. Median overall survival was 19 months. Three patients died of sepsis during remission induction, but all of other deaths were due to resistant or relapsed leukemia. The four patients who completed 3.5 year's modified L-10M protocol were free from relapse for 6-11 months (mean 8.7). Although further follow-up is necessary, we suggest that modified L-10M protocol is effective for adult ALL and long-term survival may be available.     

Intensive chemotherapy with idarubicin, cytarabine, etoposide, and G-CSF priming in patients with advanced myelodysplastic syndrome and high-risk acute myeloid leukemia

In an attempt to improve the complete remission (CR) rates and to prolong the remission duration especially in elderly patients >50 years of age, we have used a combination chemotherapy of idarubicin (10 mg/m² IV×3 days), cytarabine (AraC, 100 mg/m² CIVI×7d), and etoposide (100 mg/m²×5 days) in combination with granulocyte colony-stimulating factor (G-CSF) priming [5 mg/kg SQ day 1 until absolute neutrophil count (ANC) recovery] for remission induction. Responding patients received two consolidation courses of idarubicin, AraC, and etoposide, followed by a late consolidation course of intermediate-dose AraC (600 mg/m² IV every 12 h×5 days) and amsacrine (60 mg/m² IV×5 days). A total of 112 patients (57 male/55 female) with a median age of 58 years (range: 22–75) have been entered and are evaluable for response: 19 refractory anemia with excess of blast cells in transformation (RAEB-T), 84 acute myeloid leukemia (AML) evolving from myelodysplastic syndrome (MDS), and 9 secondary AML after chemotherapy/radiotherapy. The overall CR rate was 62%, partial remission (PR) rate 10%, treatment failure 16%, and early death rate 12%. The CR rate was higher in patients 60 years (68 vs 55%), mainly due to a lower early death rate (5 vs 21%, p<0.001). After a median follow-up of 58 months, the median overall survival is 14.5% and median duration of relapse-free survival 8 months. After 60 months, the probability of CR patients to still be in CR and alive is 16% (20% in patients 60 years and 13% in patients >60 years), while the probability of overall survival is 12% (15% in patients 60 years and 9% in patients >60 years). Compared to our previous trial (AML-MDS Study 01-92) which was done with identical chemotherapy but no G-CSF priming in 110 patients with RAEB-T, AML after MDS, or secondary AML (identical median age, age range, and distribution of subtypes), the CR rate in all patients, as well as CR rate, overall survival, and relapse-free survival in patients >60 years have significantly been improved. Thus, intensive chemotherapy with G-CSF priming is both well tolerated and highly effective for remission induction in these high-risk patients.     

The expression of co-stimulatory molecules and their relation- ship to the prognosis of human acute myeloid leukaemia: poor prognosis of B7-2-positive leukaemia

We examined the expression of co-stimulatory molecules on leukaemic cells of 52 adult patients with acute myeloid leukaemia (AML) (34 men and 18 women) and analysed the relationship between these expressions and the patient's prognosis. B7-1 was not expressed in any of the 23 patients investigated, whereas B7-2 was expressed in 26/52 patients (50.0%). B7-2 was expressed in all AML patients with monocytic morphology (M4 or M5) and in 16/42 cases without monocytic morphology. CD54 was expressed in 28/37 patients examined (75.7%), and CD58 was expressed in all of the AML patients except one (M7). The overall survival of the 26 B7-2-positive leukaemia patients (1–24 months, median survival 11.5 months) was significantly shorter than that of the 26 B7-2-negative leukaemia patients (1–71+ months, median 35.1 months) (P = 0.0080). In addition, the B7-2-positive patients exhibited significantly shorter disease-free survival periods compared to the B7-2-negative patients (P = 0.021). There was no significant difference in age, sex, haematological data and complete remission rate between the B7-2-positive and B7-2-negative patients. Our results indicated that B7-2 is one of the most crucial factors in the prognosis of adult acute leukaemia and can be expected to have an important role in tumour immunity.     

A randomized trial comparing standard versus high-dose daunorubicin induction in patients with acute myeloid leukemia

We conducted a phase 3 randomized trial comparing 2 different doses of daunorubicin as induction chemotherapy in young adults (60 years of age or younger) with acute myeloid leukemia (AML). Of 383 patients who were analyzed, 189 received standard-dose daunorubicin (SD-DN, 45 mg/m2 per day times 3 days) and 194 received high-dose daunorubicin (HD-DN, 90 mg/m2 per day times 3 days) in addition to cytarabine (200 mg/m2 per day times 7 days) to induce complete remission (CR). The CR rates were 72.0% in the SD-DN arm and 82.5% in the HD-DN arm (P = .014). At a median follow-up of 52.6 months, overall (OS) and event-free (EFS) survival were higher in the HD-DN arm than in the SD-DN arm (OS, 46.8% vs 34.6%, P = .030; EFS, 40.8% vs 28.4%, P = .030). Differences in CR rate and both OS and EFS remained significant after adjusting for other variables (CR, hazard ratio [HR], 1.802, P = .024; OS, HR, 0.739, P = .032; EFS, HR, 0.774, P = .048). The survival benefits of HD-DN therapy were evident principally in patients with intermediate-risk cytogenetic features. The toxicity profiles were similar in the 2 arms. In conclusion, HD-DN improved both the CR rate and survival duration compared with SD-DN in young adults with AML. This study is registered at www.clinicaltrials.gov as #NCT00474006.     

Significance of trilineage myelodysplasia in de novo acute myeloid leukaemia during remission rather than at diagnosis

Patients with trilineage myelodysplasia (TMDS) in de novo acute myeloid leukaemia (AML) at diagnosis and remission were clinically evaluated between 1983 and 1996. AML with TMDS (AML/TMDS) was observed in 20 (12%) of 162 patients with de novo AML at diagnosis. Complete remission (CR) was achieved with combination chemotherapy in 12 (67%) of 18 AML/TMDS cases. This CR rate was relatively worse than the rate of 78% (106/136 cases) of AML without TMDS, but this difference was not significant. Disease-free survival curves also showed no difference between AML/TMDS and AML without TMDS. During remission, eight (67%) of 12 AML/TMDS cases had myelodysplastic remission marrow (AML/MRM). AML/MRM was also seen in seven (7%) of 106 AML cases without TMDS. The actuarial disease-free survival was significantly lower in AML/MRM than in AML without MRM ( P  = 0.0003). All of the AML/MRM cases exhibited early leukaemic relapse; median remission duration was only 9 months. Clonal changes occurred in two cases of AML/TMDS and five cases of AML/MRM at the time of relapse. These findings suggest that TMDS during remission predicts a poorer prognosis and early leukaemic relapse when compared with the absence of TMDS.     

Serum levels of sCD137 (4-1BB) ligand are prognostic factors for progression in acute myeloid leukemia but not in non-Hodgkin's lymphoma

CD178 (Fas/APO-1 ligand) and CD137 ligand (CD137L) have previously been described in sera of patients with various malignancies and play an important role in the pathogenesis of various diseases. Recently, we demonstrated that low levels of soluble (s) CD137L and high levels of sCD178 correlate significantly with a long progression free survival in patients with myelodysplastic syndrome (MDS). In this study, we correlated sCD137L and sCD178 levels in sera of 42 samples of patients with acute myeloid leukemia (AML) and 46 samples of patients with non-Hodgkin's lymphoma (NHL) with stages, subtypes, and the clinical course of the diseases and determined cut-off values with maximum probability for significant differentiation between cases with higher/lower probability for progress free survival. In contrast to patients with MDS, surprisingly no correlation between sCD178 levels and different subtypes and stages or with prognosis in AML or NHL were observed. Regarding sCD137L, NHL-patients displayed lower levels compared with AML. Statistically significant higher median levels of sCD137L are present in patients with undifferentiated AML (M1/M2, 1,470 pg/mL), poor cytogenetic risk (288 pg/mL) and higher levels of BM-blasts (186 pg/mL) compared with patients with monocytoid AML (M4/M5, 89 pg/mL), intermediate cytogenetic risk (59 pg/mL) and lower levels of BM-blasts (14 pg/mL) respectively. Furthermore, in AML patients sCD137L levels correlate significantly with the probabilities to achieve complete remission (CR), stay in CR or with progress of the disease. Taken together, our data demonstrate that sCD137L can be used as a prognostic factor not only in MDS but also in AML.     

急性髓细胞白血病免疫表型分析

目的：分析成人初治急性髓细胞白血病（AML）髓系抗原、CD34及Pgp表达特点及其与预后的关系。方法：采用流式细胞仪免疫荧光标记法检测65例初治成人AML的免疫表型。结果（1）髓系抗原阳性率依次为CD33＞CD15＞CD13＞CD14,CD33阳性率高达98．46％。CD14在M5、M6表达率较高;（2）CD34及HLA－DR阳性表达率以M1、M5较高,M3最低;（3）P－糖蛋白（PG)阳性率58．33％,Pgp阳性AML的CR率28．57％,明显低于Pgp阴性的CR率（63．63％,P＜0．005）。结论AML的免疫表型检测髓系单抗可选择CD33、CD15和（或）CD14。CD34和HLA－DR低表达为M3的特征。Pgp表达与CR率有密切的关系。     

MDR 1 expression is an independent prognostic factor for response and survival in de novo acute myeloid leukaemia

The Multidrug Resistance gene (MDR 1) is frequently expressed in acute myeloid leukaemia (AML). MDR 1 is associated with resistance to chemotherapy in vitro and with a poor response rate in AML. We have investigated the prognostic value of MDR 1 expression in relation to other patient characteristics with respect to response and survival.
One hundred and thirty patients aged 0–88 years were treated for de novo AML with standard induction and consolidation chemotherapy. MDR 1 expression was determined by immunocytochemistry. Univariate and multivariate analyses were conducted to identify prognostic factors for reaching complete remission (CR) and for overall survival from diagnosis, in order to compare MDR 1 with known prognostic factors. Univariate analysis showed that higher MDR 1 expression was an adverse prognostic factor for CR ( P  < 0.001), as was higher age ( P  < 0.001) and unfavourable karyotype ( P  < 0.01). These factors were also negative prognostic factors for overall survival ( P  < 0.001, P  < 0.05 and P  < 0.005, respectively). In the multivariate analysis MDR 1 ( P  < 0.001), higher age ( P  < 0.001) and karyotype ( P  < 0.01) were independent adverse prognostic factors for CR as well as for overall survival ( P  < 0.001, P  < 0.005, P  < 0.001, respectively). Our data indicate that MDR 1 expression is a disease-related unfavourable prognostic factor which has a significant impact on complete remission and overall survival in AML. Analysis of MDR 1 may be used to determine prognosis in individual patients.     

Treatment of newly-diagnosed acute myelogenous leukaemia in patients aged 80 years and above

In order to assess outcome following treatment of acute myeloid leukaemia (AML) in patients aged 80 years and above, we have studied 33 patients aged ≥80 years treated between 1980 and 1994; 29 of these received treatment. The median age was 82 years (range 80–89). Three patients received daunorubicin (≥60mg/m² daily × 3) alone or with low-dose ara-C, two patients received '3+7' with post-treatment GM-CSF; 24 patients had higher doses of ara-C, generally with anthracyclines or fludarabine, and in nine cases with G or GM-CSF.
The median survival of the treated patients was 3–4 weeks and only two were alive after 1 year (at 66 and 79 weeks). Complete remission (CR) occurred in 9/29 (31%). Only one of the nine remains alive in remission, at 76 weeks after the date of CR, whereas the other eight died in remission or had disease recurrence at a median of 11 weeks (range 5–37 weeks) after CR. The median survival of the four untreated patients was 10 weeks (range 3–38). Patients aged ≥80 had, on average, worse outcomes than those observed in patients aged 70–79.
Our results confirm that currently available chemotherapy is generally not indicated in patients aged 80 or over with AML.     

Factors affecting the outcome of allogeneic bone marrow transplantation for adult patients with refractory or relapsed acute leukaemia

We evaluated the outcome of allogeneic bone marrow transplantation (BMT) for advanced acute myeloid leukaemia (AML) and acute lymphoblastic leukaemia (ALL) in 383 adult patients in nine Australian adult BMT centres between 1981 and 1997. The median overall survival for the group was 4.8 months, with an estimated 5-year survival of 18%. 28% of patients died of transplant-related toxicities within the first 100 d. Progressive disease was responsible for 48% of deaths. Multi-factor analysis demonstrated that AML (v ALL), disease status (second complete remission [CR2] v others), age (< 40 years) and duration of prior first complete remission (CR1) (> 6 months) were pre-transplant variables significantly associated with improved survival. Acute graft-versus-host disease (GVHD) of any grade reduced the rate of relapse in both AML and ALL, but only grades I-II were associated with improved survival. Both limited and extensive chronic GVHD were associated with increased survival. Only patients with AML in untreated first relapse or CR2, with a duration of CR1 > 6 months, or patients with T ALL, had a 5-year survival > 20%. Transplants for AML in induction failure or pre-B ALL in untreated first relapse or CR2 had an intermediate outcome, with 5-year survival of 10-20%. A 5-year survival of < 10% was observed for patients transplanted for ALL in induction failure or for pre-B ALL or AML in refractory first relapse or beyond CR2. These results suggest that for most adult patients with advanced acute leukaemia an allograft offers only a small chance of cure.     

Acute myeloid leukaemia in childhood: clinical features and prognosis

Clinical and laboratory features at presentation were correlated with morphological (FAB) subclass of AML in a group of 112 children diagnosed between 1972 and 1982. Patients with a monocytic component of AML (M4, M5) had higher initial leucocyte counts, a higher incidence of extramedullary infiltration and of CNS involvement. In M4 AML CNS relapse occurred in patients with a high initial leucocyte count whereas in M5 AML CNS involvement tended to occur at presentation in children with low initial counts. Two-thirds of patients treated achieved remission and most failures were due to inadequate chemotherapy, although haemorrhage, leucostasis or metabolic complications caused early death in patients with M4 and M5 AML. With a minimum follow up of 3 years only 12% of patients are alive; these figures have not improved in consecutive series despite increasing intensity of induction and more recent availability of bone marrow transplantation. No features predictive of long-term survival were identified, but patients with myeloid differentiation (M1, M2, M3) did better than those with a monocytic or erythroid component (M4, M5, M6). The proportion of patients with AML curable by chemotherapy seems unlikely to increase without marked intensification of post-remission chemotherapy. More aggressive CNS prophylaxis may be of benefit in cases with a monocytic component.     

Poor survival in t(8;21)(q22;q22)-associated acute myeloid leukaemia with leukocytosis

Abstract: Twenty-nine consecutive cases with a t(8;21)(q22;q22) in the bone marrow (BM) karyotype were retrospectively studied concerning clinical, morphological and cytogenetic data. All had been diagnosed as acute myeloid leukaemia (AML), 27 FAB subtype M2 and two M1, comprising 5% of all cytogenetically analysed AML during 18 yr. Auer rods were the most consistent t(8;21)-associated morphological finding and were demonstrated in 92% of the reviewed BM specimens, whereas BM eosinophilia was seen in only 24%. The median age was 53 yr, and 30% of the patients were >60 yr old. Twenty-four patients had received induction chemotherapy; 22 of these (91%) entered a complete remission (CR). The median survival time in treated patients was 18 months. Leukocytosis at diagnosis (>20×10⁹/l) was significantly (p=0.01) associated with shorter survival time. All four children are still in first CR after 9–80 months. Seven cases (25%) developed granulocytic sarcomas, discovered either at diagnosis (n=4) or at first relapse (n = 3). Secondary chromosome abnormalities were found in 62% of the cases, most often loss of a sex chromosome. The presence of such secondary aberrations did not correlate with any morphological or clinical characteristics, including survival. This first Scandinavian study of AML with t(8;21) corroborates the previous findings that these AMLs are characterized by distinct morphological features, a high frequency of CR and a striking tendency to develop extramedullary leukaemic manifestations. Leukocytosis at diagnosis indicates a less favourable prognosis.     

P-Glycoprotein Expression in Acute Myeloid Leukaemia Cells at Diagnosis: Its relationship to Daunorubicin or Idarubicin Induction Therapy and Survival; Malignancy

We investigated the expression of P-glycoprotein (P-gp) in 50 adults with de novo diagnosed acute myeloid leukaemia (AML) and the relationship between presence of P-gp in leukaemic cells and efficacy, as remission induction and survival rate, of two different anthracyclines, daunorubicin (DNR) and idarubicin (IDR). We found that 30 out of 50 patients (60%) were negative (Group 1) and 20 (40%) were positive (Group 2) for P-gp expression evaluated by mean of MRK16 MoAb using a cut-off of 10% positive cells. Thirty-five out of 50 patients (70%) obtained complete remission (CR); depending on P-gp expression, the CR rate was 80% for group 1 and 45% for group 2 (p < 0.005). The median duration of overall survival was 20 months for patients in Group 1 as compared with 10 months for patients of Group 2 (p < 0.005). Regarding the anthracycline used, no significant difference in CR was observed in patients of Group 1 (75% of CR with DNR vs. 90% with IDR); Group 2 obtained 40% of CR with DNR vs. 70% with IDR (p < 0.005). The median duration of overall survival (OS) with the two regimens was comparable in Group 1, while it was significantly longer in patients of Group 2 treated with IDR compared with DNR regimen (p < 0.005). These results confirm the prognostic value of P-gp expression in AML at first appearance and we suggest that idarubicin could be a valid anthracycline drug in the treatment of AML to be evaluated as potential drug of choice in patients with primary or drug-induced multidrug resistance.     

Relapse risk and survival in patients with FLT3 mutated acute myeloid leukemia undergoing stem cell transplantation

In patients with AML with FMS‐like tyrosine kinase 3 (FLT3) mutations, the significance of minimal residual disease (MRD) detected by PCR before allogeneic stem cell transplantation (SCT) on outcomes after transplant remains unclear. We identified 200 patients with FLT3‐AML who underwent SCT at our institution. Disease status at transplant was: first or second complete remission (CR1/CR2, n = 119), high‐risk CR (third or subsequent CR, marrow hypoplasia, or incomplete count recovery) (CR‐HR, n = 31), and morphological evidence of active disease (AD, n = 50). The median follow‐up was 27 months, and the 2‐year overall and progression‐free survival were 43% and 41%, respectively. Relapse was highest in the AD group (85%) and the CR‐HR FLT3 MRD positive group (72%), followed by CR‐HR FLT3 MRD negative (58%), CR1/CR2 FLT3 MRD positive (39%), and lowest in the CR1/CR2 FLT3 MRD negative group (23%). On multivariate analysis, independent factors influencing the risk of relapse were detectable morphological disease and FLT3 MRD by PCR pre‐transplant. Factors that did not influence the relapse risk included: age, graft type, graft source, type of FLT3 mutation, or conditioning intensity. Morphologic and molecular remission status at the time of transplant were key predictors of disease relapse and survival in patients with FLT3‐AML.     

Liposomal daunorubicin versus standard daunorubicin: long term follow-up of the GIMEMA GSI 103 AMLE randomized trial in patients older than 60 years with acute myelogenous leukaemia

This randomized phase III clinical trial explored the efficacy of DaunoXome (DNX) versus Daunorubicin (DNR) in acute myeloid leukaemia (AML) patients aged >60 years. Three hundred and one AML patients were randomized to receive DNR (45 mg/m(2) days 1-3) or DNX (80 mg/m(2) days 1-3) plus cytarabine (AraC; 100 mg/m(2) days 1-7). Patients in complete remission (CR) received a course of the same drugs as consolidation and then were randomized for maintenance with AraC+ all trans retinoic acid or no further treatment. Among 153 patients in the DNR arm, 78 (51.0%) achieved CR, 55 (35.9%) were resistant and 20 (13.1%) died during induction. Among 148 patients in the DNX arm, 73 (49.3%) achieved CR, 47 (31.8%) were resistant and 28 (18.9%) died during induction. Univariate analysis showed no difference as to induction results. After CR, DNX showed a higher incidence of early deaths (12.5% vs. 2.6% at 6 months, P = 0.053) but a lower incidence of relapse beyond 6 months (59% vs. 78% at 24 months, P = 0.064), with a cross in overall survival (OS) and disease-free survival (DFS) curves and a later advantage for DNX arm after 12 months from diagnosis. DNX seems to improve OS and DFS in the long-term follow-up, because of a reduction in late relapses.     

Idarubicin, cytarabine, and topotecan in patients with refractory or relapsed acute myelogenous leukemia and high-risk myelodysplastic syndrome.

In an effort to develop more effective therapy for patients with refractory or relapsed acute myelogenous leukemia (AML) and high-risk myelodysplastic syndrome (MDS), we investigated the efficacy of a combination chemotherapy consisting of idarubicin, cytarabine, and topotecan. Twenty-seven patients were treated: four with primary refractory AML, nine with AML in first relapse, four with AML in second relapse, and 10 with MDS-RAEB/RAEBT. Patients received as salvage therapy a single course of idarubicin 12 mg/m(2) IV bolus on days 1-3, cytarabine 1 g/m(2) over two hours q 12 hr on days 1-5, and topotecan 1.25 mg/m(2) over 24 hr on days 1-5. Median age was 42 years (range 17-65 years). All patients were evaluable for response: 14 (51.9%) achieved complete remission, 10 with AML (59%) and four with MDS (40%), respectively. Thirteen AML patients (excluding four relapsed after autologous stem cell transplantation) were grouped into four categories to stratify the probability of achieving complete remission (CR): group 1, first CR duration > or = 2 years and receiving first salvage treatment (S1); group 2, first CR duration 1-2 years and receiving S1; group 3, first CR duration 0-1 years and receiving S1; and group 4, first CR duration 0-1 years and receiving S2, S3, or S4 after failing S1. The response rate of each group was as follows: group 1, one of two (50%); group 2, one of one (100%); group 3, four of four (100%); group 4, two of six (33.3%). The median remission duration and survival of patients with AML were six and 12 months, respectively. Median duration of survival in 10 MDS patients was 15 months, and all four MDS patients achieving a CR maintained continuous CR with a median follow-up of 11 months. Severe myelosuppression was observed in all patients, resulting in fever or documented infections in 89% of patients. Median time to recovery of neutrophils > or =0.5 x 10(9)/l was 22 days (11-34) and for platelets > 20 x 10(9)/l 35 days (11-58). Reversible grade 3-4 toxicities included diarrhea (two patients) and mucositis (seven patients). We conclude that combination chemotherapy with intermediate dose cytarabine, idarubicin, and topotecan has significant antileukemic activity and acceptable toxicity in salvage AML and high-risk MDS.