ORIGINAL ARTICLE
Steroid Withdrawal Syndrome During Steroid Tapering
in Childhood Acute Lymphoblastic Leukemia
A Controlled Study Comparing Prednisone Versus
Dexamethasone in Induction Phase
Paola Saracco, MD,* Nicoletta Bertorello, MD,* Loredana Farinasso, MD,*
Silvia Einaudi, MD,† Elena Barisone, MD,* Franco Altare, BS,‡ and Guido Pastore, MD§
Summary: Children with acute lymphoblastic leukemia (ALL) re-
ceive as part of induction therapy a 4-week course of high-dose gluco-
corticoid, which is either abruptly discontinued or shortly tapered. The
aim of this study was to evaluate the signs and symptoms of steroid
withdrawal syndrome and performance status (according to Lansky
scale) during the 9-day tapering period and 1 week after withdrawal of
the steroid in 63 children randomly allocated to receive prednisone or
dexamethasone as part of induction treatment according the AIEOP
ALL 2000 protocol. Twenty of 28 (75%) patients on dexamethasone
versus 18 of 35 (51.4%) on prednisone (P , 0.05) developed at least
one steroid withdrawal symptom during the study period. Three or
more symptoms were observed in 39.3% (11/28) of the dexameth-
asone group and 8.6% (3/35) of the prednisone group (P , 0.05).
Dexamethasone patients developed clinical signs earlier (within 3 days
from the steroid tapering) than symptomatic prednisone patients. In
the prednisone group, the symptoms were less severe and the perfor-
mance status was higher (P , 0.05). Steroid withdrawal morbidity in
ALL children during induction is a frequent and clinically relevant
complaint. A more gradual (for dexamethasone) or a more prolonged
(for prednisone) tapering might be suggested.
Key Words: steroid withdrawal syndrome, acute lymphoblastic leu-
kemia, induction therapy, prednisone, dexamethasone
(J Pediatr Hematol Oncol 2005;27:141–144)
P atients receiving glucocorticoid treatment in supraphysio-
logic doses for more than 7 days develop prolonged
suppression of the hypothalamo-pituitary-adrenal (HPA) axis.
Some experience a steroid withdrawal syndrome after
discontinuation or dosage reduction of the drug even without
Received for publication November 27, 2004; accepted January 4, 2005.
From the *Department of Pediatric Onco-Hematology, University of Turin,
Turin, Italy; †Department of Pediatric Endocrinology, Regina Margherita
Children Hospital, Turin, Italy; ‡Clinical Chemistry Laboratory, Regina
Margherita Children Hospital, Turin, Italy; and §Childhood Cancer
Registry of Piedmont, Cancer Epidemiology Unit of the Centre for Cancer
Epidemiology and Prevention/CPO Piemonte, S. Giovanni Hospital,
Turin, Italy.
Reprints: Guido Pastore, Childhood Cancer Registry of Piedmont,
V. Santena 7, 10126 Torino, Italy (e-mail: pastoreguido@tin.it).
Copyright Ó 2005 by Lippincott Williams & Wilkins
J Pediatr Hematol Oncol  Volume 27, Number 3, March 2005
biochemical evidence of deficient adrenal secretion.1–3 The
most prominent symptoms of adrenal insufficiency are an-
orexia, lethargy, malaise, nausea, weight loss, headache, skin
desquamation, and fever; arthralgias, myalgias, abdominal
pain, and vomiting are less common. Postural hypotension,
hyponatremia, and hyperkalemia occur only occasionally, as
the adrenal glomerulosa is always intact.1 Grade and duration
of symptoms are extremely variable; symptoms may occur
when the steroid is being either withdrawn or reduced, even if
the tapering dosage remains supraphysiologic.2–4
Because abrupt reduction of long-term high-dose steroid
therapy is dangerous, a well-designed withdrawal schedule
must be considered in children receiving long-term steroid
therapy. Appropriate steroid tapering is related to total dose,
duration of therapy, and type of glucocorticoid used (short-
acting as hydrocortisone, intermediate-acting as prednisone,
long-acting as dexamethasone). Prolonged dexamethasone
therapy may induce more severe signs or symptoms than pred-
nisone. So far, no prospective studies have defined optimal
methods of dosage reduction in pediatric or adult patients.
Children with acute lymphoblastic leukemia (ALL)
receive, as part of induction therapy, a 4-week course of high-
dose glucocorticoid; in most worldwide protocols this is
discontinued either abruptly or after a short tapering period.5,6
Until now, few studies have been performed to assess the fre-
quency and clinical relevance of signs and symptoms related to
steroid withdrawal. A high frequency and a long persistence of
adrenal suppression have been recently reported, so a more
gradual steroid tapering schedule during ALL induction
therapy has been stressed.6,7 The aim of this prospective single-
institution study was to assess the clinical relevance of the
steroid withdrawal syndrome in children randomly allocated to
receive prednisone or dexamethasone as part of induction
treatment according to the Associazione Italiana di Ematologia
ed Oncologia Pediatrica (AIEOP) ALL 2000 protocol.8
METHODS
The study included 63 consecutive children with ALL
(32 boys, 31 girls, mean age 5.3 years [range 1.1–15.8]) ran-
domized to receive dexamethasone (28 cases) or prednisone
(35 cases), according to the AIEOP ALL 2000 protocol.8
The induction therapy consisted of two phases. During the
first treatment period children received steroid, vincristine
141
Saracco et al J Pediatr Hematol Oncol  Volume 27, Number 3, March 2005

(1.5 mg/m2, on days 8, 15, 22, 29), E. coli L-asparaginase were 6.8% and 12.2% at serum concentrations of 24.4 mg/dL
(5,000 U/m2, on days 12, 15, 18, 21, 24, 27, 30, 33), and 3.6 mg/dL, respectively (normal serum levels of cortisol:
daunorubicin (30 mg/m2 on days 8, 15, 22, 29), and three female: ,10 years, 4–23 mg/dL; 10–20 years, 6–30 mg/dL;
doses of intrathecal (IT) methotrexate. During the follow- male: ,15 years, 6–23 mg/dL; 15–20 years, 7–30 mg/dL;
ing phase, three drugs were administered: 6-mercaptopurine normal serum levels of ACTH, 10–70 pg/mL).
(60 mg/m2 from days +36 to +64), cyclophosphamide (two Informed written consent was obtained from the parents,
doses, 1,000 mg/m2 on days 36 and 64), ara-cytosine according to protocol requirements approved by the local ethi-
(75 mg/m2 from day +38 to +41, +45 to +48, +52 to +55, cal committee (including the assessment of any drug-related
and +59 to +62), and two doses of IT methotrexate. Soon after toxicity and hormonal measurements). The study design, ac-
diagnostic procedures were completed, all patients received cording to AIEOP ALL 2000 protocol guiding rules, envis-
prednisone therapy (60 mg/m2), divided in three daily doses aged no variation in the steroid tapering schedule unless
from day +1 to +7. The children were randomized either to required by severity of symptoms.
receive dexamethasone 10 mg/m2 or prednisone 60 mg/m2,
both divided into three daily doses. The full-dose steroid course Statistical Analysis
lasted 3 weeks (from day +8 to +28). From day +29 over 9 days, Frequency of withdrawal syndrome, duration of hospi-
the total dose of both steroids was tapered to 50% every 3 days talization, number of symptomatic children with a Lansky
until complete withdrawal (day +37). At the time of the study, performance score lower than 50, as well as the mean and
the 63 children were in clinical and morphologic remission and standard deviation (SD) of ACTH and cortisol levels at each
had completed the induction phase, receiving full-dose steroid testing time, were compared between the two groups using the
treatment. chi-square test or t test as appropriate. The power of the study
At day +28, during the steroid tapering (day +29, day has been calculated on observed relevant figures (proportion of
+33, days +36 to +38), and 1 week after the withdrawal of the cases or mean and SD, as appropriate).
steroid (days +41 to +48), all patients were carefully assessed
by medical staff for the occurrence of cushingoid features,
performance status, and steroid withdrawal signs and symp- RESULTS
toms. All data were prospectively collected by medical staff on At least one steroid withdrawal symptom during steroid
an ad-hoc form during controls (on the days reported above) tapering developed in 21 of 28 (75%) patients on dexameth-
and also recorded by mothers at home. Clinical manifestations asone and 18 of 35 (51.4%) on prednisone (P , 0.05). Three
were divided into subjective (eg, anorexia, lethargy, malaise, or more symptoms developed in 39.3% (11/28) of the
hyposthenia, nausea, headache, arthralgias/myalgias, abdom- dexamethasone group and in 8.6% (3/35) of the prednisone
inal pain) and objective (eg, weight loss, fever, hypotension, group (P , 0.05). Twenty of the 21 (95.2%) symptomatic
muscular hypotrophy, desquamation of the skin). A scaled dexamethasone patients developed clinical signs within 3 days
scoring system was not used due to the highly variable grading from the beginning of the steroid tapering. Among the 18
of symptoms among children; rather, a ‘‘yes/no’’ score system symptomatic prednisone patients, symptoms developed more
was employed. Due to the nonspecific clinical signs or symp- gradually: within 3 days from the beginning of tapering in 6
toms of steroid withdrawal syndrome, clinical data that might (33.3%), at the end of tapering in 9 (50%), and 3 to 6 days after
be imputable to other causes were excluded from the analysis. complete steroid discontinuation in 33 (16.7%) (Table 1).
Duration of hospitalization, number of days with a neu- Cushingoid features occurred with the same frequency in both
trophil count less than 500, and infections requiring hospital groups of patients by the end of steroid treatment.
admission and antibiotic therapy during the induction therapy
(days +1 to +64) were also recorded. Infection was considered
severe in the presence of culture-documented bacteriemia or of TABLE 1. Steroid Withdrawal Signs and Symptoms in
signs and symptoms of sepsis or other evidence of infection Children With Acute Lymphoblastic Leukemia
(eg, pneumonia, abscess, meningitis). PDN DXM P Value
Performance status was defined according to the Lansky Number of patients 35 28
scale,9 because most patients were young and it was used for Males/females 17/18 15/13
grading subjective symptoms. Mean age (yrs) 5.6 5.3
Basal corticotropin (ACTH) and cortisol levels were Number of patients
assessed. The tests were performed at diagnosis (before with withdrawal syndrome 51.4% (18/35) 75% (21/28) ,0.05
starting prednisone therapy) and 24 hours after the last dose of Number of patients
with .3 symptoms
steroid (which was given as a single dose in the morning). or symptoms .5 days 8.6% (3/35) 39.3% (11/28) ,0.05
Peripheral blood samples were collected between 8 and 9 AM Occurrence of
after overnight fasting. ACTH and cortisol levels were symptoms
measured by chemiluminescence assay with the Advantage days 1–3 from tapering 33.3% (6/18) 95.2% (20/21) ,0.05
ACTH (catalogue #62-7004, Byk Diagnostics) and Advantage Occurrence of symptoms
Cortisol (catalogue #62-7005), respectively. The interassay days 4–9 from tapering 50.0% (9/18) 4.7% (1/21) ,0.05
coefficients of variation for ACTH were 6.9% and 8.4% at Lansky ,50 during tapering 11.1% (2/18) 33.0% (7/21) ,0.05
serum concentrations of 211 pg/mL and 5.8 pg/mL, re- PDN, prednisone; DXM, dexamethasone.
spectively. The interassay coefficients of variation for cortisol

142 q 2005 Lippincott Williams & Wilkins

J Pediatr Hematol Oncol  Volume 27, Number 3, March 2005
In both groups, the majority of children reported subjective
signs and symptoms (malaise, hyposthenia, arthralgias/myalgias,
lethargy, and anorexia). The objective signs (weight loss, hypo-
tension and skin desquamation) were mainly observed in the
dexamethasone group (12 vs. 2 in the prednisone group). Two
children in the dexamethasone group required stress doses of
hydrocortisone because of severe hypotension, hyposthenia,
and malaise, occurring 2 days from the beginning of steroid
tapering.
On day +28 (the day before tapering), none of the 63
patients had a Lansky performance score less than 50. During
the steroid tapering, Lansky score less than 50 was reported in
7 of 21 symptomatic patients (33%) in the dexamethasone
group and in 2 of 18 symptomatic patients (11.1%) in the
prednisone group (P , 0.05).
The number (mean and SD) of days of severe
neutropenia and hospitalization during induction therapy were
similar between the groups (23.3 6 10.59 and 22.7 6 13.9 in
the dexamethasone group; 24.6 6 13.2 and 17.3 6 8.79 in the
prednisone group). Infectious episodes occurred in 33 of 63
patients (52.3%): 8 were severe, occurring either during the
tapering period or after steroid withdrawal (4 in the prednisone
group, 4 in the dexamethasone group).
At diagnosis, mean levels of cortisol (in the dexameth-
asone group, 14.25 mg/dL 6 9.11; in the prednisone group,
16.67 mg/dL 610.01) and of ACTH (in the dexameth-
asone group, 20.47 pg/mL 6 31.22; in the prednisone group,
13.96 pg/mL 6 14.35) were in the normal range for all
children. When steroid treatment was stopped, mean basal levels
of cortisol (5.47 mg/dL 6 5.69 in the prednisone group;
4.98 mg/dL 6 8.63 in the dexamethasone group) and of ACTH
(14.28 pg/mL 6 15.91 in the prednisone group; 10.48 pg/mL 6
11.16 in the dexamethasone group) were significantly reduced
compared with those at diagnosis (P , 0.05), without difference
between the groups. HPA axis suppression was observed in the
majority of patients: cortisol levels below 3 mg/dL (cutoff for
glucocorticoid-induced adrenal insufficiency) were reported in
62.9% (22/35) of the prednisone group and in 71.4% (20/28) of
the dexamethasone group.
The sample size of this study was adequate to detect
a difference between the two groups for the principal end point
(clinical occurrence of the steroid withdrawal syndrome)
(power 0.80, alpha 0.05). For the other minor end points, the
power was around 0.50 (alpha 0.05).
DISCUSSION
The most comprehensive and accepted definition of
steroid withdrawal syndrome is an objective syndrome
resembling true adrenal insufficiency and characterized by
fever, anorexia, nausea, lethargy, malaise, arthralgias, desqua-
mation of the skin, weakness, and weight loss occurring, with
highly variable grading, in patients undergoing steroid with-
drawal also in the presence of biochemical evidence of HPA
system integrity. Patients receiving prolonged high-dose gluco-
corticoids may experience symptoms of steroid withdrawal not
only if steroid therapy is abruptly discontinued, but also during
a rapid tapering. Steroid withdrawal syndrome can develop
even if the reduced dosage remains supraphysiologic, as dur-
q 2005 Lippincott Williams & Wilkins
Steroid Withdrawal Syndrome in ALL
ing the tapering reduction schedule. Although the mechanism
of this steroid withdrawal syndrome is not known, it is most
likely that long-term pharmacologic glucocorticoid therapy
inhibits transcription of the genes for glucocorticoid receptors,
reducing the number of receptors per cell.1–5 Thus, patients
may develop the syndrome even if they continue to receive
physiologic or even supraphysiologic doses of glucocorticoids,
as these concentrations can elicit only subphysiologic cellular
responses. These patients develop a relative state of gluco-
corticoid resistance.1,10–12 Elevated circulating concentrations
of interleukin-6 may cause many of the symptoms of gluco-
corticoid withdrawal.13
Synthetic steroids have a greater potency than cortisol.
Prednisone has four times the glucocorticoid potency of
cortisol. The glucocorticoid potency of dexamethasone is still
a matter of debate, but the anti-inflammatory effect is 30 times
stronger than the hydrocortisone effect and the ACTH
suppressive effect is 80 times stronger.14 In the present study,
the grade of withdrawal syndrome during steroid reduction
was assessed in ALL children receiving two different steroids
in the induction phase, both with a hydrocortisone equivalent
dosage of more than 200 mg/m2 per day for 28 consecutive
days. Even if a tapering reduction schedule was used, most
symptoms already occurred during the 9 days of tapering. The
schedule dose was equivalent to 10 times (in the first 3 days of
taper) to 3 times (day 9) the physiologic hydrocortisone re-
quirement. Our data showed that the steroid tapering schedule
used is not appropriate, particularly for patients on dexameth-
asone. In the dexamethasone group, symptoms were more
frequent and more severe and disabling and occurred earlier
than in the prednisone group, soon after the first taper. Patients
on prednisone developed withdrawal symptoms at a lower
dose of the steroid. This different behavior might be partly ex-
plained by the different ACTH suppressive effects of dexa-
methasone and prednisone. The mean cortisol levels at
diagnosis were in the normal range for both groups. At the
time of complete steroid suspension, after the short-term taper-
ing schedule, biochemical evidence of adrenal suppression
was present in most patients, without a significant difference
between the two groups. Recent data reported that high-dose
dexamethasone therapy for ALL is related to adrenal insuf-
ficiency lasting more than 4 to 8 weeks after discontinuation of
treatment6 and up to 4 to 8 months in a few cases,7 but clinical
data related to steroid withdrawal morbidity were lacking. A
gradual reduction of the steroid has been suggested, making
a proposal of a double-blind trial of tapering dose to determine
whether such treatment ameliorates the withdrawal syndrome
morbidity.8
The present study shows that a 9-day steroid tapering
schedule has not been adequate to avoid the withdrawal
syndrome morbidity in children treated on the AIEOP ALL
2000 protocol at a single institution. A different schedule for
the two different steroids might be suggested, and not only
for ALL children: an initially more gradual dose tapering for
dexamethasone and probably a more prolonged tapering for
prednisone. Potential downsides to tapering (long or short)
must also be considered carefully, taking into account the
increased risk of infection, increased risk of long-term bone or
CNS toxicity, or the potential for leukemia cells to develop
143
Saracco et al
steroid resistance. However, children with ALL not only are at
high risk for increased stress conditions but experience a poor
quality of life during induction treatment, and all efforts must
be made to reduce this disabling therapy-related morbidity.
When steroid reduction is associated with severe symptoms, it
might be indicated to return the glucocorticoid to its former
level and then to taper the dosage more gradually, over a more
prolonged period. Further prospective randomized studies are
needed to evaluate the impact of different steroid tapering
(more gradual for dexamethasone and more prolonged for
prednisone) and/or hydrocortisone substitution schedules on
survival in children with ALL.
REFERENCES
1. Baxter JD, Rousseau GG. Glucocorticoid hormone action: an overview.
In: Baxter JD, Rousseau GG, eds. Glucocorticoid Hormone Action. New
York: Springer-Verlag. 1979:1.
2. Dixon RB, Christy NP. On the various forms of corticosteroid withdrawal
syndrome. Am J Med. 1980;68:224–230.
3. Krasner AS. Glucocorticoid induced adrenal insufficiency. JAMA.
1999;282:671–676.
4. Morris HG, Jorgensen JR. Recovery of endogenous pituitary-adrenal
function in corticosteroid-treated children. J Pediatr. 1971;79:480–488.
144
J Pediatr Hematol Oncol  Volume 27, Number 3, March 2005
5. Schrappe M, Camitta B, Pui C-H. Long-term results of large prospective
trials in childhood acute lymphoblastic leukemia. Leukemia. 2000;14:
2193–2194.
6. Felner EI, Thompson MT, Ratliff AF, et al. Time course of recovery of
adrenal function in children treated for leukemia. J Pediatr. 2000;137:
21–24.
7. Petersen KB, Müller J, Rasmussen M, et al. Impaired adrenal function
after glucocorticoid therapy in children with acute lymphoblastic
leukemia. Med Pediatr Oncol. 2003;41:110–114.
8. Schrappe M. Protocol and Committee Progress Reports of the 13th
Annual Meeting of the International BFM Study Group (AIEOP LLA
2000 – ALL-BFM 2000); May 3–5, 2002; Budapest, Hungary.
9. Lansky SB, List MA, Lansky LL, et al. The measurement of performance
in childhood cancer patients. Cancer. 1985;60:1651–1656.
10. Miller WL. The adrenal cortex and its disorders. In: Brook CGD,
Hindmarsh PC, eds. Clinical Pediatric Endocrinology, 4th ed. Oxford:
Blackwell Science. 2001:362.
11. Tyrrel JB, Baxter JD. Glucocorticoid therapy. In: Felig P, Baxter JD,
Broadus AE, Frohman LA, eds. Endocrinology and Metabolism, 2d ed.
New York: McGraw-Hill. 1981:812–813.
12. Brook CGD. Side effects of steroids revisited. J Pediatr. 2000;137:3–4.
13. Papanicolau DA, Tsigos C, Oldfield EH, et al. Acute glucocorticoid
deficiency is associated with plasma elevations of interleukin-6: does the
latter participate in the symptomatology of the steroid withdrawal syn-
drome and adrenal insufficiency? J Clin Endocrinol Metab. 1996;81:
2303–2306.
14. Young MC, Hughes IA. In: Grossman A, ed. Clinical Endocrinology.
Oxford: Blackwell Scientific Publications. 1992:434.
q 2005 Lippincott Williams & Wilkins