Normal spontaneous cortisol secretion in children afterautologous bone marrow transplantation

PER FRISK, JAN GUSTAFSSON & JOHAN ARVIDSON

Department of Women’s and Children’s Health, Uppsala University Children’s Hospital, University Hospital,

Uppsala, Sweden

AbstractAim: To describe spontaneous cortisol secretion in children after autologous bone marrow transplantation (BMT) for acuteleukaemia and lymphoma. Methods: Spontaneous cortisol secretion was analysed in 39 children before and after BMT.Thirteen patients were conditioned with chemotherapy only (group 1), and 26 patients also with total body irradiation (TBI).In the TBI group, 14 patients had received no additional irradiation (group 2), whereas 12 patients had received cranialirradiation (CRT) previously (group 3). Results: Before BMT, in comparison with group 1, mean morning cortisol wassignificantly lower in group 2 (252 vs 415 mmol/l, p=0.004), but not in group 3 (vs 312 mmol/l, p=0.12). There was nochange in group 1 six months after BMT (to 379 nmol/l), whereas morning cortisol increased significantly in group 2 andgroup 3 (to 386 and 343 nmol/l, respectively; p50.05). The change in mean morning cortisol correlated negatively withpretransplant morning cortisol (r=70.63, p50.001). Neither TBI nor CRT were associated with changes in morningcortisol.

Conclusion: Spontaneous cortisol secretion is maintained after BMT irrespective of whether cranial or total bodyirradiation has been given or not.

Key Words: Cortisol secretion, autologous, bone marrow transplantation, children, long-term follow-up.

Introduction

Bone marrow transplantation (BMT) has evolved into

an important treatment for haematological malig-

nancies that have failed to respond to conventional

therapy. In order to eradicate remaining malignant

cell clones and to create space for the stem cells to

be infused, the child undergoes intensive condition-

ing with high doses of cytostatic agents and, when

necessary, with total body irradiation (TBI). This

intensive treatment, which is often superimposed on

preceding primary and relapse therapies, including

cranial radiation therapy (CRT), may result in multiple

endocrinological sequelae, such as hypothyroidism,

stunted growth and impaired pubertal development

[1]. Whether adrenocortical function is also affected

by BMT, and, if so, to what extent, is uncertain.

Results from previous investigations in BMT recip-

ients, most of which are cross-sectional studies in

subjects who have undergone allogeneic BMT, are

conflicting [2–5].

Mechanisms of potential impairment to the hypo-

thalamic–pituitary–adrenal (HPA) axis in the BMT

recipient include radiation-induced damage from CRT

and TBI, and the influence of high doses of corticos-

teroids given before BMT to induce remission, as well

as after BMT, above all to treat graft-versus-host

disease (GVHD). The absence of chronic GVHD in

autografted patients, which obviates the need for post-

transplant corticosteroids, may permit the study of

adrenocortical function after BMT without this

confounder. We present a longitudinal study on short-

and long-term development of spontaneous cortisol

secretion in a group of children treated with autologous

BMT for acute leukaemia and lymphoma.

Patients and methods

Between October 1985 and August 1997, 50 patients

younger than 18 y of age were treated with autologous

BMT (n=49) or syngeneic BMT (n=1) at the

University Hospital of Uppsala. Ten patients relapsed

within 6 mo after BMT. One long-term survivor was

not tested with regard to cortisol secretion. Here, we

include the 39 patients who were regularly followed

up for at least 6 mo after BMT. Baseline data are

Correspondence: Per Frisk, Akademiska Barnsjukhuset, SE-751 85 Uppsala, Sweden. Tel: +46 18 611000. Fax: +46 18 665853. E-mail: per.frisk@kbh.uu.se

(Received 10 December 2004; revised 10 February 2005; accepted 15 March 2005)

Acta Pædiatrica, 2005; 94: 1411–1415

ISSN 0803-5253 print/ISSN 1651-2227 online # 2005 Taylor & Francis Group Ltd

DOI: 10.1080/08035250510036741

summarized in Table I. Prior to BMT, patients with

acute lymphoblastic leukaemia (ALL) and lympho-

blastic lymphoma (LBL) received treatment according

to the high-risk arm of the current ALL protocol, which

included prednisolone 60 mg/m2 for 4 wk, which was

then tapered for 1 wk. If remission was ascertained,

three to four consolidation courses were given with

3–4-wk intervals, which included dexamethasone

10 mg/m2 for 5 d. Details of the conditioning regimens

have been thoroughly reported elsewhere [6]. In brief,

patients with acute myelogeneous leukaemia received

busulphan and cyclophosphamide, whereas patients

with large-cell anaplastic lymphoma and Hodgkin’s

disease received 1–3 bis chlorethyl-1 nitrosurea,

etoposide, cytarabin and cyclophosphamide. Patients

with ALL and LBL were conditioned with pred-

nisolone 100 mg/m2 for 2 d, teniposide, daunorubicin,

vincristine, cyclophosphamide and cytarabin, plus

TBI. In most patients (n=22), TBI was given in a

single fraction on day 1 as two opposed 5-MV X-ray

anterior-posterior fields with lung shielding. The total

absorbed dose in the centre of the patient was 7.5 Gray

(Gy) (dose rate 15 cGy/min), and the maximum dose

to the kidneys was also 7.5 Gy +5%. The four patients

most recently undergoing transplantation were treated

with fractionated TBI, consisting of 12 Gy in six frac-

tions over 3 d (dose rate 15 cGy/min). In addition to

TBI, 12 children had received CRT at doses ranging

from 18 to 26 Gy. The children were divided into

three groups according to radiation therapy. Group 1

consisted of those who had received no irradiation

(n=13), group 2 of those who had received TBI but

no other irradiation (n=14), and group 3 of those

who, in addition to TBI, had received CRT in the

treatment of their primary disease (n=12). Of the

four patients who had received fractionated TBI,

three patients fell into group 2 and one patient into

group 3. Nine patients relapsed during follow-up, with

a median time to relapse of 15 mo (range 7 to 48 mo).

Cortisol secretion was scheduled to be measured

before BMT (except for the first six children in the

series), 6 and 12 mo after BMT, and then annually

for at least 5 y after BMT. Median follow-up was 3 y

(range 6 mo to 5 y). No child was lost to follow-up.

Owing to procedural difficulties, not all patients

were tested at every follow-up, however. Figure 1 and

Table I show the exact number of children tested at

each point in time.

Definitions

Regular measurements using standard laboratory

methods included morning serum cortisol (06.00–

09.00 h) and evening serum cortisol (20.00–24.00 h).

The reference range at our laboratory was 250–

750 nmol/l for morning cortisol and 30–300 nmol/l for

evening cortisol. Cortisol secretion in healthy children

is unrelated to age, sex, growth or pubertal develop-

ment [7]. Hypocortisolism was defined as morning

cortisol levels less than 100 nmol/l [8].

Statistics

In agreement with previous publications, cortisol data

are reported as mean values. All other data are re-

ported as median values. One-way ANOVA was used

Table I. Baseline characteristics and morning cortisol values before and 6 mo after autologous BMT.

Diagnosis

Age

Median (range)

Before BMT

mean, nmol/l

(95% CI)

6 mo after BMT

mean, nmol/l

(95% CI)

D-cortisol

mean, nmol/l

(95% CI)

All patients 336 (289–382) 369 (318–419) 46 (714 to 107)

Group 1 AML (8),

HD (3),

LCAL (2)

13.6 (1.9–17.9) 415 (357–474)

n=13

379 (304–454)

n=12

733 (7138 to 72)

n=12

Group 2 ALL (12),

LBL (2)

7.3 (3.6–14.2) 252 (180–323)

n=10

386 (283–488)

n=11

114 (1–227)

n=8

Group 3 ALL (11),

LBL (1)

9.2 (5.6–17.7) 312 (199–424)

n=8

343 (230–455)

n=12

98 (3–193)

n=8

AML: acute myelogenous leukaemia; HD: Hodgkin’s disease; LCAL: large-cell anaplastic lymphoma; ALL: acute lymphoblastic leukaemia;

LBL: lymphoblastic lymphoma.

Figure 1. Morning and evening cortisol values including reference

values at our centre.

1412 P. Frisk et al.

to compare pretransplant values between groups and

the Bonferroni procedure to adjust for multiple

comparisons. The paired-samples t-test was used to

compare pretransplant values with those obtained after

BMT. Here, in order to avoid the problems inherent

in multiple significance testing, statistical analyses

were restricted to the short-term analysis comparing

pretransplant values with those obtained 6 mo after

BMT. Pearson’s product moment correlation was

calculated in analysis of a relationship between pre-

transplant value and the change in morning cortisol

secretion 6 mo after BMT (D-cortisol), and Spear-

man’s rank correlation in analysis of a relationship

between age at BMT and D-cortisol. Multiple regres-

sion analysis was performed to examine the contri-

bution of factors pertinent to D-cortisol. Patients

were excluded from statistical analysis when receiv-

ing corticosteroid treatment. The significance level

was set at 5%. All calculations were made with SPSS

11.0.

Ethics

This study was approved by the Ethics Committee of

the Medical Faculty, Uppsala University.

Results

Individual observations

Before BMT, no patient had a morning cortisol

level less than 100 nmol/l. After BMT, two patients,

aged 5.2 and 9.7 y, had morning cortisol values

transiently less than 100 nmol/l (44 and 98 nmol/l 6

and 24 mo after BMT, respectively); one of whom is

described below. Three patients received prolonged

corticosteroid therapy after BMT. One girl in group 1

received prednisolone due to idiopathic pneumonia

syndrome; initially 50 mg, which was gradually

tapered over 7 mo. Her subsequent cortisol values

were normal. One boy in group 2 complained of

fatigue and had a subnormal morning cortisol level

(150 nmol/l) but a normal evening cortisol level

(200 nmol/l) 2 y after BMT. Replacement therapy

(hydrocortisone 7.5 mg twice daily) resulted in an

improved sense of well-being. One girl in group 2 (with

no cortisol values available before BMT) had a low

morning cortisol level (44 nmol/l) but a normal

evening cortisol value (36 nmol/l) 6 mo after BMT

and received hydrocortisone 10 mg once daily for

3 mo. Cortisol levels were then elevated at each

follow-up visit and at the latest evaluation, 10 y after

BMT, her morning and evening cortisol levels were

901 and 409 nmol/l, respectively. These values were

interpreted to be stress-induced since she was scared

of blood sampling.

Morning cortisol

The overall mean morning cortisol before BMT in

the total population was 336 nmol/l (95% CI 289 to

382) and remained normal during follow-up (Table I

and Figure 1). On the group level, in comparison with

group 1, mean morning cortisol was significantly

lower in group 2 ( p=0.004) but not in group 3

( p=0.12) (Table I). There was no significant differ-

ence between group 2 and group 3 ( p=0.77). Six

months after BMT, there was no significant short-term

change in group 1 ( p=0.50), whereas morning cortisol

increased significantly in group 2 ( p=0.048) and in

group 3 ( p=0.045).

Including all groups, the change in the overall

mean morning cortisol level (D-cortisol) correlated

negatively with pretransplant morning cortisol

(r=70.63, p50.001), but not with age at BMT

(rs=0.046, p=0.81). In a multivariate analysis

(including the variables pretransplant morning corti-

sol, age at BMT, TBI, and CRT), only pretransplant

morning cortisol was significantly associated with

D-cortisol (B=70.61, 95% CI 71.1 to 70.05, p=0.01). Morning cortisol remained within the reference

range in all three groups during long-term follow-up

(Figure 2).

Evening cortisol

The overall mean evening cortisol before BMT was

88 (95% CI 58 to 119) nmol/l and 96 (64 to 129)

nmol/l 6 mo after BMT. Figure 1 shows that evening

cortisol appeared to develop approximately in parallel

with morning cortisol, suggesting an unchanged

circadian rhythm after BMT.

Discussion

To the best of our knowledge, this is the first lon-

gitudinal study of spontaneous cortisol secretion in

Figure 2. Morning cortisol values in groups 1–3.

Cortisol secretion and bone marrow transplantation 1413

children treated with autologous BMT. In a recent

review of late effects after BMT, Socie et al. stated that

TBI together with chronic GVHD and its treatment

constituted the major risk factors for non-malignant

complications after BMT [9]. The absence of GVHD

in autografted patients permits the elucidation of the

independent role of radiation-induced damage to

the HPA axis in BMT recipients.

We found normal levels of basal cortisol and a

preserved circadian cortisol rhythm in children up to

5 y after BMT, irrespective of whether cranial or

total body irradiation had been given or not. Individual

values were transiently subnormal in two patients after

BMT.

Before BMT the mean morning cortisol level

was lowest in the irradiated groups, albeit within the

reference range. This might be due to the prolonged

treatment given before BMT in ALL and LBL

patients, most of whom were transplanted in second

remission. The relapse therapy included repeated

courses of corticosteroids administered until shortly

before BMT, which may cause adrenal insufficiency.

Adrenal function has been shown to recover within

the first 1–2 mo after corticosteroid treatment in

ALL patients, and a sustained recovery was also

observed in the bone marrow recipients after BMT,

despite irradiation to the HPA axis [10]. The multiple

regression analysis pointed to the same conclusion,

i.e. those in whom morning cortisol was most sup-

pressed before BMT showed the greatest recovery

after BMT, whereas neither TBI nor CRT influenced

the post-transplant recovery.

In the long term, morning cortisol remained within

the reference range in all three groups. Young children

are generally more susceptible to the effects of BMT,

but this was not found in the present study, possibly

due to the small sample size [11,12].

We did not perform dynamic testing of the HPA

axis in our patients. However, as recently pointed out

in a review of late endocrine effects after childhood

cancer, disturbances in cortisol secretion may not be

reflected in dynamic testing after cranial irradiation;

a phenomenon which may represent neurosecretory

dysfunction attributable to defective hypothalamic

CRH secretion [13]. Thus, Tsatsoulis et al. reported

that six adult patients who had received cranial irra-

diation in the dose range of 20 to 40 Gy for pituitary

tumours had reduced basal cortisol secretion but

normal cortisol response to insulin-induced hypo-

glycaemia [14]. Crowne et al. observed no impairment

in spontaneous adrenocorticotrophin (ACTH) and

cortisol secretion in 20 long-term survivors of paedia-

tric ALL after conventional CRT (18 Gy) [15]. These

data suggest that neurosecretory dysfunction may be

a dose-dependent phenomenon which appears follow-

ing fractionated cranial irradiation in total doses in

excess of 20 Gy.

Results regarding the independent effect of TBI on

cortisol secretion differ between BMT reports. Most

of these have been performed cross-sectionally in

recipients of allogeneic bone marrow, and conse-

quently corticosteroids used to alleviate GVHD may

be an important confounder. Also, since cortisol is

cleared through the liver, the serum cortisol concen-

tration might be influenced by liver dysfunction, which

is more common after allogeneic BMT due to GVHD,

which may both damage the liver per se and predispose

to hepatic infections [16]. In a previous study of the

present cohort, long-term liver function was found

to be normal [17]. Sanders et al. reported that 24%

of 78 children, all of whom had received TBI, had

subnormal 11-deoxycortisol levels after metyrapone

stimulation 1–8 y after BMT, irrespective of prior

CRT, indicating a decreased pituitary ACTH reserve

[2]. In contrast with Sanders et al., but in agreement

with our data, Ogilvy-Stuart et al. showed that only

two out of 31 children who had received TBI had a

low spontaneous cortisol level, and only one child had

borderline peak cortisol response to hypoglycaemia

[3]. Other investigators have presented similar findings

[4,5]. The cause of this discrepancy is not evident but

may be related to differences in testing modalities as

well as in GVHD and its therapy. In the present study,

neither those children who had received TBI only,

nor those who had received additional CRT (and

consequently cranial irradiation doses in excess of

20 Gy) developed disturbances in spontaneous cortisol

secretion.

In conclusion, spontaneous cortisol secretion was

normal in children after autologous bone marrow

transplantation during an observation period up to 5 y

after BMT. Since a later decrease cannot be excluded,

continued long-term follow-up is necessary.

Acknowledgements

This study was supported by the Children’s Cancer Foun-dation in Sweden.

References

[1] Brennan BM, Shalet SM. Endocrine late effects after bone

marrow transplant. Br J Haematol 2002;118:58–66.

[2] Sanders JE, Pritchard S, Mahoney P, Amos D, Buckner CD,

Witherspoon RP, et al. Growth and development following

marrow transplantation for leukemia. Blood 1986;68:1129–35.

[3] Ogilvy-Stuart AL, Clark DJ, Wallace WH, Gibson BE, Stevens

RF, Shalet SM, et al. Endocrine deficit after fractionated total

body irradiation. Arch Dis Child 1992;67:1107–10.

[4] Thomas BC, Stanhope R, Plowman PN, Leiper AD. Endocrine

function following single fraction and fractionated total body

irradiation for bone marrow transplantation in childhood. Acta

Endocrinol (Copenh) 1993;128:508–12.

[5] Clement-De Boers A, Oostdijk W, Van Weel-Sipman MH,

Van den Broeck J, Wit JM, Vossen JM. Final height and

1414 P. Frisk et al.

hormonal function after bone marrow transplantation in

children. J Pediatr 1996;129:544–50.

[6] Lonnerholm G, Simonsson B, Arvidson J, Bengtsson M,

Carlson K, Hagberg H, et al. Autologous bone marrow trans-

plantation in children with acute lymphoblastic leukemia.

Acta Paediatr 1992;81:1017–22.

[7] Knutsson U, Dahlgren J, Marcus C, Rosberg S, Bronnegard M,

Stierna P, et al. Circadian cortisol rhythms in healthy boys

and girls: relationship with age, growth, body composition,

and pubertal development. J Clin Endocrinol Metab 1997;

82:536–40.

[8] Arlt W, Allolio B. Adrenal insufficiency. Lancet 2003;

361:1881–93.

[9] Socie G, Salooja N, Cohen A, Rovelli A, Carreras E, Locasciulli

A, et al. Nonmalignant late effects after allogeneic stem cell

transplantation. Blood 2003;101:3373–85.

[10] Felner EI, Thompson MT, Ratliff AF, White PC, Dickson BA.

Time course of recovery of adrenal function in children treated

for leukemia. J Pediatr 2000;137:21–4.

[11] Leiper AD. Non-endocrine late complications of bone

marrow transplantation in childhood: part I. Br J Haematol

2002;118:3–22.

[12] Leiper AD. Non-endocrine late complications of bone

marrow transplantation in childhood: part II. Br J Haematol

2002;118:23–43.

[13] Darzy KH, Gleeson HK, Shalet SM. Growth and neuro-

endocrine consequences. In: Late effects of childhood cancer.

1st ed. Published in London by Arnold; 2004. p 189–211.

[14] Tsatsoulis A, Shalet SM, Harrison J, Ratcliffe WA, Beardwell

CG, Robinson EL. Adrenocorticotrophin (ACTH) deficiency

undetected by standard dynamic tests of the hypothalamic-

pituitary-adrenal axis. Clin Endocrinol (Oxf) 1988;28:225–32.

[15] Crowne EC, Wallace WH, Gibson S, Moore CM, White A,

Shalet SM. Adrenocorticotrophin and cortisol secretion in

children after low dose cranial irradiation. Clin Endocrinol

(Oxf) 1993;39:297–305.

[16] Locasciulli A, Testa M, Valsecchi MG, Vecchi L, Longoni D,

Sparano P, et al. Morbidity and mortality due to liver disease

in children undergoing allogeneic bone marrow transplanta-

tion: a 10-year prospective study. Blood 1997;90:3799–805.

[17] Frisk P, Lonnerholm G, Oberg G. Disease of the liver following

bone marrow transplantation in children: incidence, clinical

course and outcome in a long-term perspective. Acta Paediatr

1998;87:579–83.

Cortisol secretion and bone marrow transplantation 1415