Patients with de novo acute myeloid leukaemia and complex karyotype aberrations show a poor prognosis despite intensive treatment: a study of 90 patients


Dr. Claudia Schoch, Department of Internal Medicine III, University Hospital Grosshadern, Ludwig-Maximilians-University of Munich, Marchioninistr. 15, 81377 München, Germany. E-mail:


The clinical significance of complex chromosome aberrations for adults with acute myeloid leukaemia (AML) was assessed in 920 patients with de novo AML who were karyotyped and treated within the German AML Cooperative Group (AMLCG) trials. Complex chromosome aberrations were defined as three or more numerical and/or structural chromosome aberrations excluding translocations t(8;21)(q22;q22), t(15;17)(q22;q11–q12) and inv(16)(p13q22). Complex chromosome anomalies were detected in 10% of all cases with a significantly higher incidence in patients geqslant R: gt-or-equal, slanted 60 years of age (17·8% vs. 7·8%, P < 0·0001). Clinical follow-up data were available for 90 patients. Forty-five patients were < 60 years of age and were randomly assigned to double induction therapy with either TAD-TAD [thioguanine, daunorubicin, cytosine arabinoside (AraC)] or TAD-HAM (high-dose AraC, mitoxantrone). Twenty-one patients achieved complete remission (CR) (47%), 20 patients (44%) were non-responders and 9% of patients died during aplasia (early death). The median overall survival (OS) was 7 months and the OS rate at 3 years was 12%. Patients receiving TAD-HAM showed a significantly higher CR rate than patients receiving TAD-TAD (56% vs. 23%, P = 0·04). Median event-free survival was less than 1 month in the TAD-TAD group and 2 months in the TAD-HAM group, respectively (P = 0·04), with a median OS of 4·5 months vs. 7·6 months (P = 0·13) and an OS after 3 years of 7·6% vs. 19·6%. Forty-five patients were geqslant R: gt-or-equal, slanted 60 years of age: 28 of these patient were treated for induction using one or two TAD courses and 17 cases received TAD-HAM with an age-adjusted reduction of the AraC dose. The CR rate was 44%, 38% were non-responders and 18% experienced early death. The median OS was 8 months and the OS rate at 3 years was 6%. In conclusion, complex chromosome aberrations in de novo AML predicted a dismal outcome, even when patients were treated with intensive chemotherapy. Patients under the age of 60 years with complex aberrant karyotypes may benefit from HAM treatment during induction. However, long-term survival rates are low and alternative treatment strategies for remission induction and consolidation are urgently needed.

Increasing insights into the biology of acute myeloid leukaemia (AML) disclose a considerable heterogeneity that may also underlie the clinical course and therapeutic perspectives. Recent efforts have concentrated on defining biological and clinical characteristics that may allow the discrimination of specific subtypes of AML, which may ultimately allow a risk-adapted stratification of therapy. Besides patient-associated parameters such as sex and age, leukaemia-associated parameters such as leucocyte count, FAB subtype, lactate dehydrogenase (LDH) and others were found to be related to treatment outcome and prognosis (Bloomfield, 1992; Haferlach, 1996; Wheatley et al, 1996; Büchner et al, 1997a; Cripe, 1997).

Among various pretreatment determinants, cytogenetics are unequivocally considered to be of major importance (Bloomfield et al, 1997; Mrózek et al, 1997) For clinical purposes, karyotype analysis enables discrimination between three major prognostic groups. A favourable outcome under currently used treatment regimens was observed in several studies in patients with t(8;21)(q22;q22), inv(16)(p13q22) or t(15;17)(q22; q11–12) (Berger et al, 1987; Keating et al, 1987; Fenaux et al, 1989; Dastugue et al, 1995; Hiddemann et al, 1995; Haferlach, 1996; Warrell, 1998). Chromosome aberrations with an unfavourable clinical course are −5/del(5q), −7/del(7q), inv(3)/t(3;3) and complex aberrant karyotype (Yunis et al, 1984; Berger et al, 1987; Fenaux et al, 1989; Fonatsch et al, 1994; Hiddemann et al, 1995; Haferlach, 1996; Grimwade et al, 1998). The remainder are assigned to an intermediate prognostic group. This group is very heterogenous because it includes patients with a normal karyotype and rare chromosome aberrations with an as yet unknown prognostic impact.

To date, only few data are available for patients with de novo AML and complex karyotype abnormalities (Yunis et al, 1984; Berger et al, 1987; Arthur et al, 1989; Fenaux et al, 1989; Stasi et al, 1993; Swansbury et al, 1994; Dastugue et al, 1995). Furthermore, the definition of complex aberrant karyotype varies between different study groups. Most commonly, it is defined as at least three cytogenetic abnormalities. The reported complete remission (CR) rates vary between 21% and 46%, with a median overall survival between 1 month and 5 months. The aim of this study was to characterize a large group of patients with de novo AML and complex karyotype aberrations and to analyse their response to intensive treatment.


Patients Patients aged > 18 years diagnosed with de novo AML who were admitted to the participating institutions of the German AML Cooperative Group between 1986 and 1996 were eligible. The diagnosis of AML was based on FAB criteria (Bennett et al, 1976, 1985). Patients with a history of myelodysplasia or other antecedent haematological disorder, cytotoxic therapy or radiotherapy were excluded.

Treatment Patients were treated according to the protocols of the German AML Cooperative Group (AMLCG). These comprise double induction therapy in patients < 60 years of age. In the AMLCG-86 trial, patients underwent initial randomization between two courses of thioguanine, cytosine arabinoside (AraC) and daunorubicin (TAD-TAD) or one course of TAD, followed by high-dose AraC and mitoxantrone as a second cycle (TAD-HAM) for remission induction. This was followed by TAD consolidation and subsequent monthly maintenance for 3 years. In the AMLCG-92 trial, all patients received TAD-HAM and TAD consolidation, and were randomly assigned to long-term maintenance vs. a second consolidation using a sequentially modified version of HAM. Patients aged 60 years and over received a second course of TAD or HAM with a reduced dose of AraC of 1·0 g/m2 only when they had an inadequate response to the first cycle. TAD consisted of cytarabine 100 mg/m2 by continuous intravenous infusion daily on d 1 and d 2, and by 30 min intravenous infusion every 12 h on d 3–8, with daunorubicin 60 mg/m2 by 30 min intravenous infusion on d 3, d 4 and d 5, and 6-thioguanine 100 mg/m2 orally every 12 h on d 3–9. HAM consisted of cytarabine 3 g/m2 by 3 h intravenous infusion every 12 h on d 1–3, with mitoxantrone 10 mg/m2 by 30 min intravenous infusion on d 3, d 4 and d 5 (Büchner et al, 1985, 1991; Hiddemann et al, 1987).

Prior to therapy, all patients gave their informed consent after having been advised about the purpose and investigational nature of the study, as well as the potential risks. The study is in accordance with the modified declaration of Helsinki and the protocols received approval from the ethics boards of the participating institutions.

Cytogenetic studies Chromosome analyses were carried out in four central cytogenetic laboratories performing 75% of cytogenetic studies and in six local laboratories. Pretreatment bone marrow or blood were analysed cytogenetically. Chromosome analyses were performed using short-term cultures according to standard protocols using G- or R-banding. The chromosomes were interpreted according to the International System for Human Cytogenetic Nomenclature (ISCN, 1995).

The failure rate was 3% to 10% in the central laboratories and higher in the local ones, leading to an overall failure rate of 13%. While cytogenetic analysis was successful in only 81% of cases in the trial performed from 1986 to 1992, the success rate increased to 93% in the trial started in 1992. After introduction of all-trans retinoic acid (ATRA) into therapy for acute promyelocytic leukaemia, patients with AML M3/t(15;17) were entered into a different trial. Cytogenetic results were available for 920 patients. Aberrant karyotypes were detected in 463 patients (49·8%).

Complex karyotype was defined by the presence of at least three clonal cytogenetic abnormalities. Patients with favourable chromosome aberrations [t(8;21)(q22;q22), t(15;17)(q22;q11–12) and inv(16)(p13q22)] and additional chromosome abnormalities were excluded because the biology of these subgroups of AML appears different.

Clinical end-points Complete remission (CR) was defined as a normocellular bone marrow containing < 5% blasts and neutrophil granulocytes > 1·5 × 109/l and platelets > 100 × 109/l in the peripheral blood, according to Cancer and Leukaemia Group B (CALGB) criteria (Preisler et al, 1979). Causes of treatment failure were subdivided into two categories: persistent leukaemia (non-response) and death within 42 d of beginning induction therapy (early death). Overall survival was measured from the time of entry into the study to the time of death. Patients still alive were considered as censored observations at the date of last follow-up. Event-free survival (EFS) was calculated from the time of therapy start to either failure to attain CR, relapse of AML after achievement of CR or death from any cause respectively. For non-responders, EFS was set at 0 months. CR duration was measured from the date of attainment of CR until relapse or death.

Statistical methods Comparison of rates of complete remission were evaluated by chi-square test. The Kaplan–Meier (1958) method was used to estimate the distribution of CR duration, event-free survival and overall survival. Comparisons of overall survival and EFS between age groups and treatment groups were performed with the log-rank test (Peto et al, 1977).


Ninety-four patients with a complex aberrant karyotype were identified among a total of 920 consecutive, previously untreated adult patients with newly diagnosed de novo AML. Complete follow-up data were available for 90 patients.

Patient characteristics

Presenting clinical and haematological features of the 90 patients with complex aberrant karyotypes are summarized in Table I. The median age was 59·5 years (range 19–81 years) with 45 patients < 60 years of age and 45 patients geqslant R: gt-or-equal, slanted 60 years of age. All FAB subtypes were observed with the exception of AML M3 and AML M4eo (owing to exclusion criteria).

Table I.  Clinical and morphological data of 90 patients with de novo acute myeloid leukaemia (AML) and complex aberrant karyotype.
 All patients< 60 years of agegeqslant R: gt-or-equal, slanted 60 years of age
Number of patients904545
Male–female ratio38:52 = 1:1·419:26 = 1:1·419:26 = 1:1·4
Age range (years)19–8119–5960–81
Median age59·54768
Leucocyte count (range, × 109/l)0·5–3500·5–3500·7–225
Median leucocyte count (× 109/l)8·012·525·75
Platelet count (range, × 109/l)3·0–3153·0–3155·0–107
Median platelet count (× 109/l)43·567·035·0
Haemoglobin concentration range (g/dl)3·8–13·83·8–13·84·6–12·1
Median haemoglobin concentration (g/dl)9·08·659·1
FAB-morphologyM0: 2M0: 2 
M1: 18M1: 9M1: 9
M2: 29M2: 11M2: 18
M4: 17M4: 11M4: 6
M5: 9M5: 5M5: 4
M6: 9M6: 3M6: 6
M7: 4M7: 2M7: 2
no data: 2no data: 2 


The overall incidence of complex karyotype aberrations was 10·2% in the study population. In patients under 60 years, the incidence (7·8%) was significantly lower than in patients 60 years and older (17·8%) (P < 0·0001).

Twenty-four patients of the total 90 patients for whom follow-up data were available showed three or four chromosome abnormalities, 43 cases showed 5 to 10 abnormalities and in 23 cases, more than 10 chromosome aberrations were observed. Nineteen (21%) patients showed neither an involvement of chromosome 5 nor of chromosome 7. Only numerical chromosome abnormalities were found in three patients. Table II gives an overview of the incidence of primary abnormalities in our patients with complex aberrant karyotype, as defined by Johansson et al (1994).

Table II.  Complex aberrant karyotypes categorized according to primary aberrations (Johansson et al, 1994).
Primary chromosome aberrationn
  • *

     Four patients showed a del(5)(q) and del(7)(q)/−7 in addition.

  •  †

     In 12 patients del(7)(q)/−7 was also observed and in five cases +8 was observed.

  •  Abnormalities were classified according to a hierarchy following the listed order so that each patient is only listed once, although many patients had aberrations fitting in more than one category.

inv(3)(q21q26)/t(3;3)(q21;q26) 5*
t(11)(q23) 2
+8 7
+13 1
+21 1
None of the above or any aberration defined as primary10

Cytogenetics and clinical outcome

The CR rate was 64% in all treated patients, while only 46% of patients with complex aberrant karyotype achieved CR (P = 0·0005). Concerning all patients treated in the AMLCG trials, the CR rate was higher in patients younger than 60 years of age than in patients of 60 years and older (69% vs. 55%, P = 0·001). With reference to patients with complex karyotype abnormalities, CR rates were comparable in younger and elderly patients (47% vs. 44%, P = 0·83). No significant differences were observed between both age groups concerning median event-free survival or median overall survival (< 1 vs. < 1 month and 7 vs. 8 months respectively) (Table III, Figs 1 and 2).

Table III.  Survival data of 90 patients with de novo acute myeloid leukaemia (AML) and complex aberrant karyotype.
All patients

< 60 years
< 60 years:
< 60 years:

geqslant R: gt-or-equal, slanted 60 years
Number of patients9045133245
Complete remission rate41/90 = 45·6%21/45 = 46·7% 3/13 = 23·1%18/32 = 56·3%20/45 = 44·4%
Non responder37/90 = 41·1%20/45 = 44·4% 9/13 = 69·2%11/32 = 34·4%17/45 = 37·8%
Early death rate12/90 = 13·3% 4/45 = 8·9% 1/13 = 7·7% 3/32 = 9·4% 8/45 = 17·8%
Relapses33/41 = 80·5%15/21 = 71·4% 3/3 = 100%12/18 = 66·7%18/20 = 90%
Median time to relapse
 (months, range)
 6 (1–40) 6 (1–30) 5 (5–13) 7 (1–30) 6 (1–40)
Median CR duration (months)78086
CCR at 3 years12%15%0%18·5%11%
Median event-free survival
< 1< 1< 12< 1
Event-free survival at 3 years5·7%7·6%0%11%4·6%
Median overall survival (months)774·57·68
Overall survival at 3 years11%11·7%7·6%19·6%5·7%
Median overall survival,
 responders only (months)
Overall survival at 3 years,
 responders only
Number of patients receiving
 a bone marrow transplantation
allo CR2
allo CR1: 2, allo CR2: 2
auto CR1: 2, auto CR2: 1,
syngen CR1: 1
Figure 1.

Event-free survival of 90 patients with AML and complex karyotype according to either age < 60 years or geqslant R: gt-or-equal, slanted 60 years. Patients alive and in remission are indicated by the short, broken, vertical lines.

Figure 2.

Overall survival of 90 patients with AML and complex karyotype according to either age < 60 years or geqslant R: gt-or-equal, slanted 60 years. Short, broken, vertical lines indicate patients alive.

The Medical Research Council (MRC) study group defined complex aberrant karyotype as five or more chromosome abnormalities (Grimwade et al, 1998). Sixty-six of our patients fulfilled this definition. Analysis of survival data did not show any difference of this subgroup compared with the whole group (data not shown).

Patients younger than 60 years of age at diagnosis

Thirteen patients received double induction therapy with two courses of TAD. Nine patients did not respond to treatment, one died as a result of infection and only three patients attained a complete remission. All three patients relapsed 5 months, 5 months and 13 months after CR respectively. One of these patients received an allogeneic bone marrow transplantation (BMT) in second CR and was alive at last follow-up 91 months after diagnosis and 71 months after BMT respectively.

Thirty-two patients were treated with TAD-HAM as the induction protocol. In 18 patients, a complete remission was achieved, 11 patients were non-responders and three died within the first 6 weeks owing to infection or bleeding. Twelve out of 18 responders relapsed 1–30 months after attaining CR.

While the early death rate was comparable in the TAD-TAD and TAD-HAM groups, patients receiving TAD-HAM showed a significantly higher CR rate (56%) than patients receiving TAD-TAD (CR rate 23%)(P = 0·04). Median EFS was less than 1 month in the TAD-TAD group and 2 months in the TAD-HAM group, respectively (P = 0·04), with a median OS of 4·5 months vs. 7·6 months (P = 0·13) (Figs 3 and 4).

Figure 3.

Event-free survival of 45 patients with AML younger than 60 years of age with complex karyotype according to induction therapy TAD-TAD vs. TAD-HAM. Short, vertical lines indicate patients alive and in remission.

Figure 4.

Overall survival of 45 patients with AML younger than 60 years of age with complex karyotype according to induction therapy TAD-TAD vs. TAD-HAM. Short, vertical lines indicate patients alive.

In nine patients, BMT was performed. Six of these patients received an allograft. One of the two patients transplanted in first complete remission (CR1) is in continuous CR (CCR) at 51+ months, while the other died owing to relapse 20 months after initial diagnosis and 11 months after BMT respectively. Two out of three patients transplanted in second CR (CR2) are alive in CR 71 months and 3 months after BMT respectively. One patient receiving an allograft from an identical twin in CR1 is alive in CCR at 39+ months after diagnosis. In three patients, an autologous BMT was performed, two in CR1 (one patient died owing to transplant-related toxicity, one patient is alive in CCR at 6+ months) and one in CR2, who is alive in CR at 12+ months after BMT.

Patients 60 years of age or older at diagnosis

Twenty-eight patients were treated for induction using one or two TAD courses. Six of these cases received 30 mg/m2 instead of 60 mg/m2 × 3 daunorubicin; 17 patients were treated with TAD-HAM. No significant differences were observed between the different treatment groups with respect to CR rate and overall survival.

Twenty out of 45 patients achieved a complete remission (44·4%), but 18 of these 20 patients relapsed within 1–40 months (median 6 months). Only two patients are alive in CR1 (92+ and 8+ months) and three patients are alive in CR2 (4+, 7+ and 10+ months).


As different therapeutic strategies have become available in the treatment of AML, the definition of distinct subgroups with different prognoses and therapeutic perspectives gains increasing importance for the appropriate selection of therapy. In recent years, the karyotype of leukaemic blasts has become the single most important prognostic determinant for AML, both for initial response to induction therapy, as well as for remission duration and overall survival (Bloomfield et al, 1997; Mrózek et al, 1997) In this study, we characterized a group of patients with de novo AML and complex karyotype abnormalities who showed a poor prognosis despite intensive treatment.

Out of 90 patients with de novo AML and complex aberrant karyotypes, only 46% achieved a complete remission, which was of short duration (median CR duration: 7·0 months), in spite of intensive double induction therapy and subsequent consolidation and maintenance in remission. These data are in accordance with other reports of low CR rates and short survival times for patients with complex karyotype aberrations and/or abnormalities of chromosomes 5 and/or 7 (Yunis et al, 1984; Berger et al, 1987; Fenaux et al, 1989; Haferlach, 1996). These aberrations are frequently associated and complex karyotypes often include −5/del(5q) and −7/del(7q) abnormalities. Recently, Ravandi et al (1997) reported that patients suffering from AML, myelodysplastic syndrome (MDS) refractory anaemia with excess of blasts (RAEB) or RAEB in transformation (RAEB-T) with simple −5/−7 abnormalities experienced a significantly better outcome than patients with −5/−7 and complex aberrant karyotype abnormalities. In two earlier studies, a rather good prognosis for patients with isolated 7q– or −7 was also observed (Fenaux et al, 1989; Swansbury et al, 1994). We also found evidence that the clinical outcome of patients with unfavourable but not complex aberrant karyotype (5/5q–, −7/7q–, 3q abnormalities, 11q23 abnormalities, 12p abnormalities or 17p abnormalities) differed from those patients with complex aberrant karyotype: patients with complex karyotype showed a significantly shorter relapse-free survival (4 months vs. 8 months, P = 0·004) and overall survival (8 months vs. 12 months, P = 0·042) (unpublished data). In our study, 19 out of 90 patients showed neither an involvement of chromosome 5 nor of chromosome 7, but clinical outcome did not differ from patients with abnormalities of chromosomes 5 and/or 7. Therefore, the complexity of the karyotype itself seems to be relevant for prognosis, independent of the involvement of chromosomes 5 and/or 7. Data from Pedersen-Bjergaard et al (1990) support this finding by demonstrating that the number of chromosome aberrations was an independent prognostic factor for patients with secondary myelodysplastic syndrome.

Complex karyotype aberrations are also frequently found in secondary AML occurring after treatment with radiotherapy and/or alkylating agents. The Groupe Francais de Cytogenetique Hematologique (1994) reported on 21 patients with treatment-related AML and complex aberrant karyotype who received intensive induction therapy. CR was achieved in 24% of cases and median CR duration was only 4 months. Although the CR rate was higher in our patients with de novo AML and complex karyotype abnormalities, the dismal outcome is similar to patients with de novo and secondary AML showing both complex karyotype aberrations.

In our study, the prognosis of younger patients with complex aberrant karyotype was as poor as in elderly cases. In contrast, Dastugue et al (1995) observed that complex aberrant karyotypes were predictive of short survival in older patients only. This discrepancy might be as a result of the fact that the younger patient group with complex aberrant karyotypes in the French study consisted of only 10 cases, three of whom had favourable chromosome abnormalities [t(8;21)(q22;q22) and inv(16)(p13q22)].

Bloomfield et al (1998) analysed the effect of dose intensification of AraC in post-remission therapy according to karyotype. Remission duration was prolonged in patients with t(8;21)(q22;q22), inv(16)(p13q22)/t(16;16)(p13;q22), del(16) or normal karyotype, but not in those with other karyotypic abnormalities. The poor overall outlook for patients with complex karyotype abnormalities is only marginally improved by including high-dose AraC into the induction treatment. Thus, a beneficial effect may only be expected from allogeneic bone marrow or peripheral stem cell transplantation. In our study, all three long-term survivors had received an allogeneic BMT. However, Gale et al (1995) pointed out that the impact of cytogenetic abnormalities may not be overcome by this procedure. Hence, leukaemia-free survival was lower in recipients of HLA-compatible sibling transplants with poor prognosis cytogenetics (24%) than in all other subgroups (54%) owing to a significantly higher rate of leukaemic recurrences (58% vs. 21%). In two further studies, Ferrant et al (1995, 1997) also observed that myeloablative therapy may not correct for the poor prognosis of an unfavourable karyotype. The main cause of failure was relapse, illustrating the inefficiency of current therapeutic approaches to eradicate the malignant clone in most patients, even if allogeneic BMT is performed. In the MRC AML 10 trial, patients with adverse cytogenetics did equally poorly after chemotherapy alone, autologous or allogeneic bone marrow transplantation (Grimwade et al, 1998).

The resulting genetic heterogeneity may also explain the high degree of treatment resistance. Hence, expression of the multidrug resistance protein mdr1 is associated with unfavourable cytogenetics (Willman, 1996; Leith et al, 1997), especially with abnormalities of chromosome 7 (Guerci et al, 1995). Data from Wattel et al (1994) also demonstrate an association of p53 mutations with unfavourable karyotypes and, in particular, with complex cytogenetic abnormalities. We detected a deletion of one p53 allele in 14 out of 30 patients with AML and complex karyotype abnormalities analysed using fluorescence in situ hybridization.

Recent studies from our own group also indicated that leukaemic blasts with complex aberrant karyotypes have a low proliferative activity that is, by itself, also associated with a poor response to induction therapy (Jahns-Streubel et al, 1997). These data may lead to new therapeutic perspectives that take the biology of AML with complex aberrant karyotypes into account. Chemotherapy priming with granulocyte colony-stimulating factor (G-CSF) and granulocyte-macrophage CSF (GM-CSF) was tested in several clinical trials. To date, however, priming in vivo with G-CSF or GM-CSF has not been shown to augment the efficacy of chemotherapy (Büchner et al, 1997b; Löwenberg et al, 1997; Terpstra & Löwenberg, 1997). This might be owing to the fact that only a small subgroup might benefit from priming strategies. Recent in vitro data from Jahns-Streubel et al (1996) demonstrated that proliferative activity of blasts with unfavourable karyotypes was as a result of low production of growth stimulatory cytokines. Adding GM-CSF to the culture medium resulted in a significant increase in proliferation and growth rate. Priming with G-CSF or GM-CSF might, therefore, be a promising concept for improving treatment results in the subgroup of patients with complex karyotype abnormalities.

In conclusion, our data show that patients with complex aberrant karyotypes account for 10% of patients with de novo AML. The outcome of these patients treated with standard intensive therapy is poor at both younger and higher ages. Prognosis is similar to patients with complex karyotype aberrations suffering from secondary AML. Therefore, new therapeutic options have to be tested in this group. Younger patients might benefit from an intensification of induction therapy, including high-dose AraC, to increase the CR rate, immediately followed by an allogeneic bone marrow transplantation. For elderly patients not suitable for intensification, it has to be discussed whether a palliative treatment may be the better approach (Hiddemann et al, 1999). Most importantly, however, new insights into the biology of these leukaemias need to be gained that may not only enable a better understanding but also a more effective treatment of this AML subtype.


We thank all participating centres of the AMLCG trial for sending material for cytogenetic analyses and documentation of clinical data. The AMLCG is supported by Deutsche Krebshilfe, Project No. M17/92/Bü1.