Leukemia (2000) 14, 1736–1742

 2000 Macmillan Publishers Ltd All rights reserved 0887-6924/00 $15.00

www.nature.com/leu

Treatment of childhood acute myelogenous leukemia with an intensive regimen

(AML-87) that individualizes etoposide and cytarabine dosages: short- and long-term
effects
MK Arnaout1, KM Radomski2, DK Srivastava3, X Tong3, JR Belt4, SC Raimondi5, FG Behm5, VM Santana6,7, WR Crom2,8,
J Mirro Jr6,7 and RC Ribeiro1,6,7

Departments of 6Hematology-Oncology, 1International Outreach Program, 2Pharmaceutical Sciences, 3Biostatistics and Epidemiology,
5
Pathology, and 4Molecular Pharmacology, St Jude Children’s Research Hospital, and 7Department of Pediatrics, University of Tennessee-
Memphis College of Medicine, Memphis, TN, USA

The purpose of this study was to assess the feasibility and
efficacy of a treatment regimen for pediatric acute myelogen-
ous leukemia (AML) that uses four rotating drug pairs and
adjusts dosages of etoposide and cytarabine to target specific
plasma concentrations. Thirty-one girls and 27 boys (median
age, 9.7 years) with de novo AML were treated on the protocol.
Six cycles of chemotherapy were planned. Cycles 1 to 4
comprised the drug combinations cytarabine plus etoposide,
cytarabine plus daunomycin, etoposide plus amsacrine, and
etoposide plus azacitidine, respectively. For cycles 5 and 6, the
first two combinations were repeated. Dosages were adjusted
to achieve plasma concentrations of 1.0 ␮M ± 0.1 ␮M cytarabine
and 30 ␮M ± 0.3 ␮M etoposide. Forty-four patients (76%) entered
complete remission. Of those, 24 have had relapses; 23 remain
alive in first or subsequent remission. The 5-year event-free
survival (EFS) estimate was 31.0% ± 5.9%; the 5-year survival
estimate was 41.4% ± 6.3%. Six patients (10%) died of the toxic
effects of therapy. Severe neutropenia occurred in all cycles.
Long-term complications of therapy included hepatitis C, car-
diac insufficiency, and hearing loss. Adjustment of cytarabine
and etoposide dosage was feasible for achieving targeted
plasma drug concentrations; however, the potential clinical
efficacy of this approach was offset by substantial acute and
long-term toxicity. Leukemia (2000) 14, 1736–1742.
Keywords: AML; etoposide; cytarabine; late effects; children; pedi-
atric
Introduction
The prospect of long-term survival for children with acute
myelogenous leukemia (AML) remains relatively poor.1
Although more than 80% of these children achieve complete
remission on contemporary intensive chemotherapy regimens,
only about 50% are likely to be cured.2,3 Primary or secondary
resistance to chemotherapy remains the most common cause
of treatment failure. A number of studies have found that
intensification of chemotherapy during induction and consoli-
dation phases improves outcomes.4–7
Cytarabine (Ara-C), anthracyclines, and etoposide (VP-16)
are often used to treat AML. However, the pharmacokinetic
disposition of these agents in children can vary extensively.8,9
Patients who clear these drugs rapidly may have inadequate
systemic exposure that results in treatment failure. Conversely,
patients who eliminate these agents slowly are at greater risk
of life-threatening toxicity. With the aim of eliminating
interpatient variability in exposure to Ara-C and VP-16, we
Correspondence: RC Ribeiro, Department of Hematology-Oncology,
St Jude Children’s Research Hospital, 332 N Lauderdale, Memphis,
TN, 38105–2794, USA; Fax: 901 495 3122
8
Current address: 4385 Sequoia Road, Memphis, TN 38117–1640,
USA
Received 22 March 2000; accepted 21 June 2000
used pharmacokinetic methods to individualize the dosages
of these drugs in each course in which they were given. Here
we address the feasibility of incorporating this individualized
dosage adjustment method into a treatment plan for children
with AML and assess the short-term and long-term conse-
quences of this strategy.
Patients and methods
Patients
Between April 1987 and March 1991, 67 patients were
enrolled on the AML-87 protocol. Fifty-eight were previously
untreated patients with de novo AML, six patients had second-
ary AML, and two patients had myelodysplastic syndrome.
The remaining patient was found upon subsequent review to
have acute lymphoblastic leukemia and was reassigned to the
appropriate treatment protocol. Standard morphologic, cyto-
chemical, and immunophenotypic10 criteria were used to
establish the diagnosis and classification of AML according
to the French–American–British (FAB) system.11 Cytogenetic
studies of all patients’ malignant cells were conducted.
The institutional review board of St Jude Children’s Research
Hospital approved all investigations. Written informed
consent was obtained from patients, parents, or guardians,
as appropriate.
Treatment plan
Remission induction therapy consisted of three cycles of
chemotherapy with four agents. In cycle 1, Ara-C was admin-
istered as a 96-h continuous subcutaneous infusion at an
initial dosage of 500 mg/m2 per day. The dosage was adjusted
as described below to achieve a targeted plasma concen-
tration. Concurrently, VP-16 was administered intravenously:
a loading dose of 70 mg/m2 given over 1 h was followed by
a 96-h infusion at an initial dosage of 500 mg/m2 per day; the
dosage was adjusted as described below. In cycle 2, Ara-C
was administered as in cycle 1, but over a period of 120 h. A
bolus infusion of daunomycin (50 mg/m2) was administered
on days 1 and 2. In cycle 3, VP-16 was administered as in
cycle 1, and amsacrine (m-AMSA; 125 mg/m2 per day) was
administered as a continuous 72-h intravenous infusion
(Figure 1). Cycles 2 and 3 were begun as soon as hemato-
poietic recovery was adequate (absolute neutrophil count
⬎0.5 × 109/l and platelet count ⬎75 × 109/l) and gastrointes-
tinal and infection complications had resolved, but were
begun without these requirements in the case of persistent
marrow infiltration or progressive disease. Consolidation ther-

Figure 1 AML-87 chemotherapy schedule showing targeted plasma concentrations of cytarabine and etoposide. m-AMSA, amsacrine; Ara-
C, cytarabine; 5-AZ, azacytidine; VP-16, etoposide.
apy consisted of three additional cycles of combination
chemotherapy. In cycle 4, VP-16 was administered as in cycle
1, and bolus intravenous infusions of 5-azacytidine (5-AZ;
300 mg/m2) were given 24 and 48 h after completion of the
VP-16 infusion. Cycle 5 was identical to cycle 2, and cycle 6
was identical to cycle 1. No maintenance chemotherapy was
used in this protocol. All-trans retinoic acid was not used for
patients with acute promyelocytic leukemia.
Patients with no evidence of central nervous system (CNS)
leukemia received five courses of triple intrathecal therapy
(methotrexate, hydrocortisone, and Ara-C (MHA)) adjusted for
age.12 One course was given with each cycle of systemic ther-
apy, starting with cycle 2. Children who had CNS involvement
at the time of diagnosis were treated with the intrathecal regi-
men weekly for at least 4 weeks or until the CSF was clear,
and then with each cycle of systemic therapy. Patients with
CNS involvement who remained in complete remission
received cranial irradiation (24 cGy) after the completion of
chemotherapy.
Complete response (CR) was defined as the complete
disappearance of measurable disease, restored bone marrow
function, and normal clinical performance status.
Complications of treatment and supportive care
Complications of treatment were detected and evaluated as
follows: hearing loss and speech delay by audiometry; cardiac
dysfunction by clinical cardiovascular examination,
diagnostic imaging (echocardiography, chest radiography), or
electrocardiography; clinical neurological disorder (seizure)
by electroencephalography or diagnostic imaging; endocrine
dysfunction by weight above the 97th percentile (obesity), and
by laboratory tests of thyroid function, serum gonadotropin,
androgen, estrogen, and hormone stimulation or suppression;
learning disabilities by neuropsychological evaluation of glo-
bal intellect and academic achievement; and cataracts by
ophthalmologic examination.
Supportive care was available as needed and included the
best medical management of hemorrhage, infection, weight
loss, or disseminated intravascular coagulation. Patients with
mucositis received platelet transfusions to maintain a platelet
count ⬎20 000/␮l. Trimethoprim-sulfamethoxazole chemo-
therapy was begun on day 14 of induction for prophylaxis
of Pneumocystis carinii pneumonitis. Fevers accompanied by
neutropenia were treated with triple broad-spectrum anti-
biotics; amphotericin B was routinely added if fever persisted.
Hematopoietic colony-stimulating factors were not used in
The AML-87 protocol for children
MK Arnaout et al
1737
this study. All patients were treated in rooms with positive-
pressure high-efficiency particulate air (HEPA) filtration units.
Pharmacokinetic studies and dosage adjustment
In cycle 1, plasma samples were obtained 1, 6, and 24 h after
the beginning of the Ara-C and VP-16 infusions for measure-
ment of plasma drug concentrations; samples were obtained
subsequently as needed for dosage adjustment. Plasma drug
concentrations were assayed by high-performance liquid
chromatography (HPLC) as previously described;13,14 ultra-
violet absorbance was used to detect Ara-C, and electro-
chemical detection was used to detect VP-16. The inter-day
coefficient of variation was consistently less than 3.6% for the
Ara-C assay and less than 7.1% for the VP-16 assay.
Samples collected at hours 1 and 6 were immediately
assayed. Dosage adjustment was based on plasma concen-
trations at hour 6. If the drug concentration fell within 10%
of the targeted concentration (0.9 to 1.1 ␮M for Ara-C and 27
to 33 ␮M for VP-16), no dosage adjustment was made. If the
plasma concentration was outside this range, the dosage was
adjusted proportionally to achieve the targeted concentration
(for example, if a VP-16 dosage of 500 mg/m2 per day pro-
duced a plasma concentration of 40 ␮M, the dosage was
adjusted to 375 mg/m2 per day to target a concentration of
30 ␮M). Assays of plasma samples obtained after hour 6 were
used to assess the efficacy of the targeting strategy. Subsequent
cycles began at the dosage last administered in the previous
cycle, and plasma samples were obtained as needed for dos-
age adjustment. Ara-C apparent clearance and VP-16 plasma
clearance rates were calculated for patients whose steady-
state plasma drug concentrations were measured after dose
adjustment. VP-16 clearance was calculated as
k0
Cl = ,
Css
where k0 is the final infusion rate and Css is the steady-state
plasma concentration. Because Ara-C was administered as a
subcutaneous,15 rather than an intravenous infusion, apparent
clearance was calculated as
k0
Cl/F = ,
Css
where F is bioavailability, k0 is the final infusion rate, and Css
is the steady-state plasma concentration.
Leukemia
 The AML-87 protocol for children
MK Arnaout et al
1738
Statistical analysis
The duration of event-free survival was measured as the inter-
val between study entry and the first adverse event (relapse,
withdrawal from the study, or death from any cause) or the
most recent follow-up examination. Patients who did not have
a CR were assigned an EFS value of zero. The duration of
survival was measured as the interval between entry on the
study and either the last follow-up examination or death. Dis-
tributions of survival rates and EFS rates were estimated by the
method of Kaplan and Meier16 and were compared by using
the Mantel–Haenszel statistic. All estimates of outcome are
reported within (±) one standard error (s.e.). Fisher’s exact test
was used to determine whether an association existed
between the rate of CR and the presence of defined risk
factors.
To assess the association between plasma drug clearance
and toxicity, we used Student’s t-test to compare the average
clearance of Ara-C and of VP-16 in two groups of patients
(those who experienced fewer than five episodes of grade 3
or 4 toxicity and those who experienced five or more
episodes). We also examined the association between these
two groups and the dosages of Ara-C (⬍500 or ⭓500 mg/m2
per day) and VP-16 (⬍500 or ⭓500 mg/m2 per day) by using
Fisher’s exact test. Student’s t-test was used to compare the
average number of episodes of grade 3 or 4 toxicity experi-
enced by patients who received ⬍500 and ⭓500 mg/m2 per
day of Ara-C and of VP-16. All comparisons were considered
significant at a type I error rate of ⬀ = 0.05, and all P values
reported reflect two-sided tests. All statistical analyses used
SAS Release 6.12 software (SAS, Cary, NC, USA).
Results
Patients
The median age at diagnosis of the 27 boys and 31 girls with
de novo AML was 9.7 years (range, 0.5 to 20.7 years). At the
time of diagnosis, the median WBC count was 13.8 × 109/l
(range, 1.0 to 249 × 109/l), the median hemoglobin concen-
tration was 82 g/l (range, 32 to 161 g/l), and the median plate-
let count was 49 × 109/l (range, 5 to 547 × 109/l). The FAB
AML subtypes, in order of frequency, were M2 (38%), M5
(19%), M3 (12%), M7 (10%), M1 (10%), M4 (8.6%), and M0
(1.7%). Cytogenetic abnormalities were present in 82.8% of
the cases: t(8;21) in 24.1%, inv(16) in 8.6%, t(15;17) in
10.3%, t(9;11) in 5.2%, other 11q23 abnormalities in 6.9%,
and miscellaneous karyotypic structural changes in 22.4% of
Table 1 Pharmacokinetics of cytarabine and VP-16 during cycle 1 of chemotherapy
Variable Ara-C
n Median Minimum
Dosage (mg/m2·24 h) 58 518.3 303.9
Plasma concentration (␮M) before 21 0.74 0.45
dosage adjustment
Steady-state concentration (␮M) 21 0.95 0.25
Clearance (l/h·m2) 21 150.8a 47.5a
a
Apparent clearance is reported for cytarabine because of its administration by subcutaneous infusion.
Ara-C, cytarabine; VP-16, etoposide.
Leukemia
cases. Numerical chromosome changes included +21 (3.4%)
and −7 (1.7%).
Pharmacokinetic studies
Fifty-eight patients were included in the pharmacokinetic
analysis (Table 1). The median Ara-C dosage required to
achieve a plasma concentration of 1 ␮M ± 0.1 ␮M was
518.3 mg/m2 (range, 303.9 to 867.5 mg/m2) per day. The
median VP-16 dosage required to achieve a plasma concen-
tration of 30 ␮M ± 0.3 ␮M was 511.2 mg/m2 (range, 362.1 to
1004.2 mg/m2) per day.
In cycle 1, the median Ara-C plasma concentration at hour
6 of infusion (before dosage adjustment) was 0.74 ␮M (range,
0.45 to 1.65 ␮M), and the median steady-state concentration
after dosage adjustment was 0.95 ␮M (range, 0.25 to 1.44 ␮M).
More than one dosage adjustment was required in six cycles.
The median Ara-C apparent clearance rate was 150.8 l/h·m2
(range, 47.5 to 791.7 l/h·m2) among the 21 patients who had
measurement of steady-state drug concentration.
As shown in Table 1, the median concentration of VP-16 at
hour 6 of the first infusion (before dosage adjustment) was
26.8 ␮M (range, 18.4 to 59.1 ␮M). The median VP-16 steady-
state concentration after dosage adjustment was 20.8 ␮M
(range, 21.1 to 39.1 ␮M). More than one dosage adjustment
was required in eight cycles. The median VP-16 clearance rate
of the 23 patients who had steady-state plasma drug
concentrations measured was 1.54 l/h·m2 (range, 0.68 to
3.17 l/h·m2).
In cycles subsequent to cycle 1, additional blood samples
were drawn from fewer patients for dosage adjustment. For
that reason, steady-state concentrations achieved in different
cycles could not be compared.
Outcome
Nine of the 67 patients enrolled on the study (six with second-
ary AML, two with MDS, and one who was misdiagnosed)
were excluded from the outcome analysis. Of the 58 patients
with de novo AML, two withdrew from the study because of
excessive toxicity; one of these patients received only one
cycle of induction chemotherapy, and one received only two
cycles. These two patients were included in the event-free sur-
vival analysis as having experienced an adverse event at the
time they withdrew from the study. However, both remain free
of disease, 9 years after the diagnosis of AML. Seventy-six per-
cent of the patients (44 of 58) had a complete response to
VP-16
Maximum n Median Minimum Maximum
867.5 58 511.2 362.1 1004.2
1.65 23 26.8 18.4 59.1
1.44 23 28.8 21.1 39.1
791.7a 23 1.54 0.68 3.17
 The AML-87 protocol for children
MK Arnaout et al

Figure 2 Survival estimates of children with de novo AML enrolled
on the AML-87 study.
therapy. Twenty-three (52%) had a CR after the first course
(Ara-C plus VP-16), 19 after the second course (Ara-C plus
daunomycin), and one after the third course (VP-16 plus m-
AMSA). One patient had a complete response only after the
fourth course (VP-16 plus 5-AZ). Of the 14 patients who did
not enter remission during treatment on the AML-87 protocol,
only one remains alive, 10 years after diagnosis; this patient
had a subsequent response to higher-dose Ara-C. Twenty-four
patients experienced hematologic relapse after achieving a
CR. Of those, five remain alive after receiving high-dose
chemotherapy followed by hematopoietic stem cell trans-
plantation (HSCT); the remaining 19 patients died after sal-
vage chemotherapy, HSCT, or both. Six patients died during
remission from complications of therapy and one died during
remission as the result of an automobile accident. At a median
follow-up of 6.4 years (range, 1 to 12 years), 23 patients
remain alive with no evidence of disease. Of the nine patients
who underwent HSCT because of resistant or recurrent dis-
ease, six remain alive. Figure 2 shows estimates of survival for
all patients enrolled on the AML87 study. The overall survival
estimate at 5 years and at 10 years was 41.4% ± 6.3% and
41.4% ± 11.2%, respectively.
Toxicity
Table 2 summarizes the severe (grade 3 or 4) toxic effects
observed in each cycle of chemotherapy. Severe myelosup-
1739
pression occurred in all of the 288 cycles of chemotherapy
administered. Myelosuppression was most severe during
cycles 1 and 6, which combined Ara-C with VP-16. Gastro-
intestinal toxicity, mainly mucositis, occurred in 136 cycles
(47.2%), and was most frequently seen in cycle 1. Severe
infectious complications (documented infections with sys-
temic inflammatory response) were noted in 108 cycles
(37.5%) and were most frequent in cycle 5. Hepatotoxicity
(increased bilirubin concentration, elevated transaminase
activity, or both) occurred in 82 cycles (28%). This compli-
cation was most frequent in cycles 1 (35%) and 2 (21%); in
the other cycles, only 7% to 15% of patients were affected.
VP-16 with m-AMSA (cycle 3) was the best-tolerated combi-
nation. Nausea and vomiting or diarrhea that necessitated
inpatient care were infrequent (11 cycles; 4%). Severe skin
manifestations, including rash, peeling, blistering, and des-
quamation, were noted in 14 cycles (5%), predominantly
cycles 1 and 6 (Ara-C plus VP-16). Neurologic toxicity
(seizures that resolved with medical treatment) occurred in
three patients. There was no association between toxicity and
the rate of clearance of VP-16 (P = 0.57) or Ara-C (P = 0.84).
Grade 3 and 4 toxicity was equally prevalent among patients
who received more than and patients who received less than
500 mg/m2 per day of Ara-C or VP-16.
Treatment-associated complications often delayed the start
of subsequent cycles of chemotherapy during induction. The
median time between the beginning of the first and second
courses was 27 days (range, 13 to 56 days), and a median of
34 days (range, 19 to 103 days) elapsed between the second
and third courses.
Patients were considered to be long-term survivors 10 years
after completion of therapy. There were 23 long-term sur-
vivors at the time of this analysis. The late complications seen
in this group are summarized in Table 3. The incidence of
endocrine abnormalities, including ovarian failure, growth
hormone deficiency, and hypothyroidism, was higher among
patients who underwent allogenic HSCT, all of whom
underwent total body irradiation as part of the preparatory
regimen. Only one patient, who underwent hematopoietic
stem cell transplantation, had symptomatic heart failure; four
patients had asymptomatic abnormally low shortening frac-
tion. Hepatitis C infection with chronic liver disease was
noted in four patients who underwent chemotherapy alone
and in one who underwent autologous HSCT. Three patients
who underwent chemotherapy alone had sensorineural hear-
ing loss. Cataracts were found in one patient who underwent
allogeneic HSCT.

Table 2 Toxic effects according to cycle of chemotherapy on the AML-87 protocol

Category Cycle 1 Cycle 2 Cycle 3 Cycle 4 Cycle 5 Cycle 6

Ara-C, Ara-C, VP-16, VP-16, Ara-C, Ara-C,
VP-16 daunomycin m-AMSA 5-AZ daunomycin VP-16
(n = 58) (n = 55) (n = 53) (n = 48) (n = 41) (n = 33)

No. % No. % No. % No. % No. % No. %

Hematopoietic 58 100 55 100 53 100 48 100 41 100 33 100

Neurologic 1 1.72 1 1.8 1 1.9 1 2.1 0 0 0 0
Hepatic 29 50 17 30.9 10 18.9 12 25 6 14.6 8 24.2
Gastrointestinal 46 79.3 14 25.5 23 43.4 26 52.1 5 12.2 22 66.7
Cutaneous 7 12.1 2 3.6 1 1.9 0 0 0 0 4 12.1
Infectious 18 31.0 19 34.6 13 24.5 20 41.7 24 58.5 14 42.4
Other 4 6.9 1 1.8 1 1.9 0 0 0 0 0 0

m-AMSA, amsacrine; Ara-C, cytarabine; 5-AZ, 5-azacytidine; VP-16, etoposide.

Leukemia
 The AML-87 protocol for children
MK Arnaout et al

1740
Table 3 Late toxic effects among 23 patients treated for de novo event-free survival estimate was only 33%. Disease recurrence
AML on the AML-87 protocol was the primary cause of treatment failure.25 German and
English investigators have used alternative early-intensification
Late effect No. patients Patients Patients approaches in which Ara-C, daunomycin, VP-16, and thio-
(%) treated with treated with
chemotherapy chemotherapy
guanine combinations are administered at variable intervals,
alone (n) and HSCT depending on the patient’s response.3,26 Rates of remission
(n) and estimates of survival and EFS were comparable to those
obtained by the most successful early intensification treat-
Auditory 3 (13) 2 1 ment approaches.2
Cardiac 5 (22) 3 2 We reasoned that some of the disparities observed in the
Endocrine results of early intensification studies were caused by variable
Hypothyroidism 1 (4) 0 1
individual drug disposition. A strategy that could ensure the
GH deficiency 1 (4) 0 1
Ovarian failure 2 (9) 0 2 achievement of a predetermined plasma drug concentration
Obesity 6 (26) 6 0 should enhance the effectiveness of early treatment intensifi-
Precocious puberty 1 (4) 1 0 cation by ensuring adequate systemic exposure for patients
Hepatitis C 5 (22) 4 1 whose rate of drug clearance is high. Such a strategy should
Learning disabilities 7 (30) 6 0 also reduce the severity of toxicity for patients whose
Neurological disorders 2 (9) 2 1
Speech dysfunction 3 (13) 2 1
clearance rates are low. Like previous studies, our analysis
Visual (cataracts) 1 (4) 0 1 demonstrated great variability among patients in the plasma
pharmacokinetics of Ara-C and VP-16;27 apparent clearance
rates of Ara-C varied by 20-fold, and clearance rates of VP-
16 by eight-fold.28 By adjusting the dosages of these drugs on
Discussion the basis of plasma concentration at hour 6 of the infusion,
we were able to achieve plasma concentrations within 10%
The AML-87 treatment protocol was designed to reduce the of the targeted values. Median steady-state plasma drug
variability of systemic drug exposure among patients, and thus concentrations were within the targeted range 6 h after dosage
to increase the efficacy of therapy, by individualizing Ara-C adjustment, a finding indicating that this strategy is feasible.
and VP-16 doses to achieve predetermined steady-state However, disease control was not improved. The complete
plasma concentrations. This approach had been shown to response rate of 76% was not superior to the rates obtained
produce more consistent dose intensity than does the use of in other trials employing alternative approaches.2,3,29,30
standardized dosage based on body surface area.9 Although It is not clear why this approach did not improve patient
no direct correlation has been found between plasma outcomes. We considered two possible explanations. First, the
concentration of Ara-C and intracellular concentration of Ara- use of dose-intensive VP-16 with Ara-C might have caused
CTP, its active metabolite, low plasma exposure may limit
severe toxicity without adding substantially to the overall
Ara-CTP accumulation,15,17 whereas high plasma levels may
treatment efficacy. Gastrointestinal toxicity or severe infec-
cause unnecessary toxicity. The targeted Ara-C plasma con-
tion, which affected almost 80% of patients during the first
centration, 1.0 ␮M, was selected on the basis of data suggest-
course of induction chemotherapy, often delayed subsequent
ing that it would result in an effective intracellular Ara-CTP
courses. However, although the estimated 5-year EFS rate was
concentration.15,18,19 The targeted VP-16 concentration,
higher for patients whose second induction course began
30 ␮M, was selected to produce a total leukemic cell and
within 30 days after the beginning of the first course
systemic exposure similar to that reported to be effective in
(35% ± 7.6%) than for other patients (25% ± 8.8%), this differ-
several studies.20,21
Severely myelosuppressive chemotherapy during the initial ence was not found to be statistically significant. The second
phase of treatment appears to be crucial for long-term disease- possibility is that the regimen acted selectively. In a previous
free survival in AML.4,22 Early intensification of induction ther- study, a regimen using targeted systemic drug exposure
apy is based on in vitro evidence that leukemic cells are improved outcomes for patients with B-lineage, but not T-lin-
recruited into the cell cycle 6 to 10 days after initial exposure eage, acute lymphoblastic leukemia.9 In our study, outcomes
to cytotoxic agents23 or on the theory (Goldie–Coldman may have differed among subsets of patients, obscuring the
hypothesis) that the initial therapy selects for resistant leu- effect of more consistent systemic exposure. However, the
kemic clones that arise by spontaneous mutation.24 These study lacked sufficient statistical power to identify prognostic
theoretical assumptions have guided several studies. Investi- factors correlated with response to therapy.
gators from the Children’s Cancer Group randomly assigned Toxicity was not related to the rate of clearance of Ara-C
patients to receive five-drug induction chemotherapy either or VP-16 or to dosages of either drug greater than 500 mg/m2
on a predetermined schedule (all drugs given on days 0–4 and per day. Individual adjustment of the Ara-C and VP-16 dosage
10–14) or on a flexible schedule (on days 0–4 and 14–18 or is likely to have minimized any extremes of plasma drug con-
later, depending on the patient’s response). Outcome was sig- centration caused by rapid or slow drug clearance. Severe
nificantly better for the patients whose chemotherapy was not myelosuppression, gastrointestinal toxicity, and rash were
delayed (3-year EFS estimate, 42% vs 27%).2 Using a similar most frequent in cycles that combined Ara-C with VP-16, a
strategy, St Jude investigators tested the value of early intensi- finding suggesting that the targeted concentrations chosen
fication of induction therapy with a regimen that combined may have been too high for this combination therapy. How-
five drugs given in three cycles: Ara-C plus VP-16; Ara-C, dau- ever, the rate of mortality during the first month of induction
nomycin, and thioguanine; and VP-16 plus 5-AZ. Successive therapy (3.4%) was lower than the rates reported for other
cycles were started immediately (on day 10 of the first cycle) regimens.2 Because fewer patients had dosage adjustments
in the absence of severe bone marrow hypoplasia. Although after cycle 1, we were unable to compare steady-state drug
the rate of remission induction was high (85%), the 2-year concentration across cycles. Steady-state concentrations in

Leukemia

later cycles may have been less well controlled than those in
cycle 1 were shown to be.
The late effects of the AML-87 treatment protocol were
assessed in patients who survived more than 10 years after
diagnosis. Among those who underwent HSCT, the long-term
sequelae were those associated with total-body irradiation.31
The high incidence of hepatitis C-induced chronic liver
disease was presumably caused by extensive blood product
support at a time when reliable hepatitis C screening was
unavailable. Obesity, a well known complication of antileuke-
mia treatment in acute lymphoblastic leukemia,32 was
observed in 30% of the survivors. Many also had hearing loss
of the sensorineural type, probably caused by ototoxic anti-
biotics used to treat infectious complications. Because of the
relatively low cumulative dose of anthracycline, symptomatic
cardiac dysfunction was not a problem among our patients.
We have demonstrated that individualized dosages of Ara-
C and VP-16 can produce more consistent systemic drug
exposure among patients, thus effectively standardizing the
systemic dose intensity of therapy. However, this multiagent
regimen caused severe toxicity, and its long-term efficacy was
inferior to some of the regimens2,3,33 that use traditional
methods of dosage determination.
Acknowledgements
This work was supported in part by NCI Cancer Center Core
Grants P30-CA-21765 and PO1-CA-20180 and by the Amer-
ican Lebanese Syrian Associated Charities (ALSAC). The
authors extend special thanks to Sharon Naron for editorial
assistance.
References
1 Pui CH. Acute leukemia in children. Curr Opin Hematol 1996; 3:
249–258.
2 Woods WG, Kobrinsky N, Buckley JD, Lee JW, Sanders J, Neudorf
S, Gold S, Barnard DR, DeSwarte J, Dusenbery K, Kalousek D,
Arthur DC, Lange BJ. Timed-sequential induction therapy
improves postremission outcome in acute myeloid leukemia: a
report from the Children’s Cancer Group. Blood 1996; 87:
4979–4989.
3 Stevens RF, Hann IM, Wheatley K, Gray RG. Marked improve-
ments in outcome with chemotherapy alone in paediatric acute
myeloid leukaemia: results of the United Kingdom Medical
Research Council’s 10th AML trial. MRC Childhood Leukaemia
Working Party. Br J Haematol 1998; 101: 130–140.
4 Mayer RJ, Davis RB, Schiffer CA, Berg DT, Powell BL, Schulman
P, Omura GA, Moore JO, McIntyre OR, Frei E III. Intensive postre-
mission chemotherapy in adults with acute myeloid leukemia.
Cancer and Leukemia Group B. New Engl J Med 1994; 331:
896–903.
5 Rees JK, Gray RG, Wheatley K. Dose intensification in acute
myeloid leukaemia: greater effectiveness at lower cost. Principal
report of the Medical Research Council’s AML9 study. MRC Leu-
kaemia in Adults Working Party. Br J Haematol 1996; 94: 89–98.
6 Bishop JF, Lowenthal RM, Joshua D, Matthews JP, Todd D,
Cobcroft R, Whiteside MG, Kronenberg H, Ma D, Dodds A. Etopo-
side in acute nonlymphocytic leukemia. Australian Leukemia
Study Group. Blood 1990; 75: 27–32.
7 Weick JK, Kopecky KJ, Appelbaum FR, Head DR, Kingsbury LL,
Balcerzak SP, Bickers JN, Hynes HE, Welborn JL, Simon SR,
Grever M. A randomized investigation of high-dose versus stan-
dard-dose cytosine arabinoside with daunorubicin in patients with
previously untreated acute myeloid leukemia: a Southwest
Oncology Group study. Blood 1996; 88: 2841–2851.
8 Evans WE, Sinkule JA, Crom WR, Dow L, Look AT, Rivera G. Phar-
macokinetics of teniposide (VM26) and etoposide (VP16–213) in
The AML-87 protocol for children
MK Arnaout et al
1741
children with cancer. Cancer Chemother Pharmacol 1982; 7:
147–150.
9 Evans WE, Relling MV, Rodman JH, Crom WR, Boyett JM, Pui CH.
Conventional compared with individualized chemotherapy for
childhood acute lymphoblastic leukemia. New Engl J Med 1998;
338: 499–505.
10 Behm FG. Diagnosis of childhood acute myeloid leukemia. Clin
Lab Med 1999; 19: 187–237.
11 Bennett JM, Catovsky D, Daniel MT, Flandrin G, Galton DA, Gral-
nick HR, Sultan C. Proposed revised criteria for the classification
of acute myeloid leukemia. A report of the French–American–Bri-
tish Cooperative Group. Ann Intern Med 1985; 103: 620–625.
12 Pui CH, Mahmoud HH, Rivera GK, Hancock ML, Sandlund JT,
Behm FG, Head DR, Relling MV, Ribeiro RC, Rubnitz JE, Kun
LE, Evans WE. Early intensification of intrathecal chemotherapy
virtually eliminates central nervous system relapse in children with
acute lymphoblastic leukemia. Blood 1998; 92: 411–415.
13 Sinkule JA, Evans WE. High-performance liquid chromatographic
assay for cytosine arabinoside, uracil arabinoside and some
related nucleosides. J Chromatogr 1983; 274: 87–93.
14 Sinkule JA, Evans WE. High-performance liquid chromatographic
analysis of the semisynthetic epipodophyllotoxins teniposide and
etoposide using electrochemical detection. J Pharm Sci 1984; 73:
164–168.
15 Liliemark JO, Paul CY, Gahrton CG, Peterson CO. Pharmacokinet-
ics of 1-beta-D-arabinofuranosylcytosine 5′-triphosphate in leu-
kemic cells after intravenous and subcutaneous administration of
1-beta-D-arabinofuranosylcytosine. Cancer Res 1985; 45:
2373–2375.
16 Kaplan EL, Meier P. Nonparametric estimation from incomplete
observations. J Am Stat Assoc 1999; 53: 457–481.
17 Plunkett W, Iacoboni S, Keating MJ. Cellular pharmacology and
optimal therapeutic concentrations of 1-beta-D-arabinofuranosyl-
cytosine 5′-triphosphate in leukemic blasts during treatment of
refractory leukemia with high-dose 1-beta-D-arabinofuranosylcy-
tosine. Scand J Haematol Suppl 1986; 44: 51–59.
18 Plunkett W, Heinemann V, Estey E, Keating M. Pharmacologically
directed design of leukemia therapy. Hamatol Bluttransfus 1990;
33: 610–613.
19 Ross DD, Thompson BW, Joneckis CC, Akman SA, Schiffer CA.
Metabolism of ara-C by blast cells from patients with ANLL. Blood
1986; 68: 76–82.
20 Hande KR, Wedlund PJ, Noone RM, Wilkinson GR, Greco FA,
Wolff SN. Pharmacokinetics of high-dose etoposide (VP-16–213)
administered to cancer patients. Cancer Res 1984; 44: 379–382.
21 Johnson DH, Hande KR, Porter LL, Hainsworth JD, Wolff SN,
Greco FA. High-dose pilot studies in extensive-stage small-cell
lung cancer. Semin Oncol 1986; 13: 37–39.
22 Bishop JF, Matthews JP, Young GA, Szer J, Gillett A, Joshua D,
Bradstock K, Enno A, Wolf MM, Fox R. A randomized study of
high-dose cytarabine in induction in acute myeloid leukemia.
Blood 1996; 87: 1710–1717.
23 Lampkin BC, Nagao T, Mauer AM. Synchronization of the mitotic
cycle in acute leukaemia. Nature 1969; 222: 1274–1275.
24 Goldie JH, Coldman AJ, Gudauskas GA. Rationale for the use of
alternating non-cross-resistant chemotherapy. Cancer Treat Rep
1982; 66: 439–449.
25 Kalwinsky D, Mirro J Jr, Schell M, Behm F, Mason C, Dahl GV.
Early intensification of chemotherapy for childhood acute nonlym-
phoblastic leukemia: improved remission induction with a five-
drug regimen including etoposide. J Clin Oncol 1988; 6: 1134–
1143.
26 Creutzig U, Ritter J, Schellong G. Identification of two risk groups
in childhood acute myelogenous leukemia after therapy intensifi-
cation in study AML-BFM-83 as compared with study AML-BFM-
78. AML-BFM Study Group. Blood 1990; 75: 1932–1940.
27 Ochs J, Rivera GK, Pollock BH, Buchanan G, Crist W, Freeman
AI. Teniposide (VM-26) and continuous infusion cytosine arabino-
side for initial induction failure in childhood acute lymphoblastic
leukemia. A Pediatric Oncology Group pilot study. Cancer 1990;
66: 1671–1677.
28 Donehower RC, Karp JE, Burke PJ. Pharmacology and toxicity of
high-dose cytarabine by 72-hour continuous infusion. Cancer
Treat Rep 1986; 70: 1059–1065.
29 Creutzig U, Ritter J, Zimmermann M, Schellong G. Does cranial
Leukemia
 The AML-87 protocol for children
MK Arnaout et al
1742
irradiation reduce the risk for bone marrow relapse in acute mye-
logenous leukemia? Unexpected results of the Childhood Acute
Myelogenous Leukemia Study BFM-87. J Clin Oncol 1993; 11:
279–286.
30 Ravindranath Y, Yeager AM, Chang MN, Steuber CP, Krischer J,
Graham-Pole J, Carroll A, Inoue S, Camitta B, Weinstein HJ. Auto-
logous bone marrow transplantation versus intensive consoli-
dation chemotherapy for acute myeloid leukemia in childhood.
Pediatric Oncology Group. New Engl J Med 1996; 334: 1428–
1434.
Leukemia
31 Robinson N, Sullivan KM. Complications of allogeneic bone mar-
row transplantation. Curr Opin Hematol 1994; 1: 406–411.
32 Didi M, Didcock E, Davies HA, Ogilvy-Stuart AL, Wales JK, Shalet
SM. High incidence of obesity in young adults after treatment of
acute lymphoblastic leukemia in childhood. J Pediatr 1995; 127:
63–67.
33 Creutzig U, Zimmermann M, Ritter J, Henze G, Graf N, Loffler H,
Schellong G. Definition of a standard-risk group in children with
AML. Br J Haematol 1999; 104: 630–639.