rir(ate

TI

rhR6ap

Gcod

prdcsaa

J

BGII

5

Traumatic Brain Injury is a Rarely Reported Cause of GrowthHormone Deficiency

ANNA MCDONALD, MRCPCH, MARGARETHA LINDELL, DAVID B. DUNGER, MD, FRCP, FRCPCH, AND

CARLO L. ACERINI, MD, FRCP, FRCPCH

We determined the frequency of traumatic brain injury (TBI)-related growth hormone deficiency (GHD) from a largeegistry of growth hormone– deficient subjects and compared these subjects’ clinical characteristics with those of children withdiopathic growth hormone deficiency (IGHD). A surprisingly small number of subjects with TBI-induced GHD (n � 141) wereegistered compared with those with IGHD (n � 23,722). At onset of treatment, the subjects with TBI-induced GHD were olderP � .045), had lower height velocity (P < .001), had a greater number of other pituitary hormone deficiencies (P < .001) and,fter a year of recombinant human GH treatment, demonstrated a greater change in height velocity (P � .016). We speculatehat TBI-induced GHD may be a neglected phenomenon in childhood, and recommend prospective longitudinal studies toxplore its natural history and frequency. (J Pediatr 2008;152:590-3)

he association between traumatic brain injury (TBI) and long-term hypothalamic pituitary dysfunction has long beenrecognized.1 Until recently, the incidence of this disorder after TBI was thought to be low. However, recent retrospective andprospective studies in adult TBI survivors suggest a prevalence of anterior pituitary deficiency ranging from 23% to 69%. 2-8

t remains unclear how the severity and mechanism of injury affects the frequency and type of pituitary defects that develop.Little data are available regarding the epidemiology and clinical presentation of hypopituitarism after TBI in children; until

ecently, data were limited to case series in which various pituitary hormone disturbances, including precocious puberty, growthormone (GH) deficiency, hypoadrenalism, hypogonadism, hypothyroidism, and cranial diabetes insipidus, were described. 9

ecently, 2 small systematic studies have reported frequencies of post-TBI hypothalamic pituitary dysfunction of 10.4% and1%, but these studies are limited in their conclusions due to either their retrospective design or their use of suboptimalssessment methods.10,11 Data on the prevalence, type, and natural history of pituitary dysfunction after TBI in the pediatricopulation remain sparse.

We used the information available from a large pharmacoepidemiologic database of children receiving recombinant humanH (rhGH) for GHD to determine the number of patients registered with TBI as the cause of GHD and to compare the

linical and endocrinologic features of patients with idiopathic GHD (IGHD) and those with GHD secondary to TBI. Ourbjective was to gain insight into the epidemiology and presenting clinical characteristics of TBI-induced hypothalamic-pituitaryysfunction in this age group.

METHODSData were retrieved from the Pfizer International Growth (KIGS) database, a large

harmacoepidemiologic database comprising anonymous data on children and adolescentseceiving therapy with rhGH starting in 1987. The database’s primary function is toocument the long-term outcome and safety of rhGH therapy, allowing centers toompare prescribing practice and treatment outcomes against national and internationaltandards. Children and adolescents are registered by their clinicians on a voluntary basisfter parental informed consent. Currently, 58 countries contribute to the KIGS database,nd at the time of the present study, approximately 62,000 patients had been registered.

Data were collected on all children registered in the database between 1987 andanuary 2006. The etiologic classification of either TBI or IGHD was determined. The

MI Body mass indexHD Growth hormone deficiency

GHD Idiopathic growth hormone deficiency

rhGH recombinant human growth hormoneSDS Standard deviation scoreTBI Traumatic brain injury

From the Department of Pediatrics, Uni-versity of Cambridge, Cambridge, UK(A.M., D.D., C.A.) and KIGS, Pfizer Inc,Stockholm, Sweden (M.L.).

D.D. and C.A. have received either re-search funding or honoraria for speaking atmeetings or serving on advisory board pan-els on behalf of Pfizer.

Submitted for publication Aug 3, 2007; lastrevision received Nov 12, 2007; acceptedDec 18, 2007.

Reprint requests: Dr Carlo L. Acerini, De-partment of Pediatrics, Level 8/Box 116,Addenbrooke’s Hospital, Cambridge CB22QQ, UK. E-mail: cla22@cam.ac.uk

0022-3476/$ - see front matter

Copyright © 2008 Mosby Inc. All rightsreserved.

TT Insulin tolerance test

90

10.1016/j.jpeds.2007.12.046

doepVticrgtbWa

dTsa8ragicg

y(

1Ahtyw6Pw9si0

dgwsclgd

tIcTmi4

T

APHHMB

N

T

IMTAGA

T

efinition of IGHD was made according to the KIGS etiol-gy classification criteria, defined as GHD in which “notiology is established.” GHD was defined on the basis of aeak GH value �10 ng/mL on a dynamic stimulation test.arious GH stimulation tests were used, with the insulin

olerance test (ITT) the most common in both groups (61.7%n the TBI group and 52.6% the IGHD group). Baselinelinical and biochemical characteristics and data on growthesponse to rhGH treatment were retrieved for both patientroups. The KIGS database contains no information onhe cause of TBI of affected subjects. The data comparisonetween the TBI and IGHD groups was analyzed using

ilcoxon’s rank-sum test. A P value of �.05 was defineds statistical significance.

RESULTSA total of 23,863 patients were identified on the KIGS

atabase with GHD due to either IGHD (n � 23,722) orBI (n � 141). Data on magnetic resonance imaging (MRI)

cans of the brain were recorded in 61.7% of the TBI patientsnd in 52.6% of the IGHD patients. In the IGHD patients,9.2% of the MRI scans were reported as normal, and in theemaining 10.8% of scans, all structural abnormalities were ofnon–hypothalamic-pituitary nature. In contrast, in the TBIroup, 38.2% of the MRI scans had demonstrable abnormal-ties of the hypothalamic or pituitary structures. The baselinelinical and biochemical characteristics for the 2 groups areiven in Table I.

The median age of initiation of rhGH treatment wasounger in the IGHD group compared with the TBI group

able I. Baseline clinical and biochemical characteris

IGHD

n Median

ge of onset of treatment, years 23,722 10.3repubertal 23,723 (87.7%)eight, SDS 23,715 �2.43eight velocity, cm/year 8947 4.43ax growth hormone, mU/L 22,447 5.95MI, SDS 23,722 �0.26

S, not significant.

able II. Pituitary hormone deficiencies

IG

n

solated GHD 22,851ultiple deficiencies 22,851SH deficiency 20,881drenocorticotropic hormone deficiency 20,421onadotrophin deficiency 20,275ntidiuretic hormone deficiency 20,212

median [10th to 90th percentile range], 10.3 years [4.4 to t

raumatic Brain Injury is a Rarely Reported Cause of Growth Hormone

4.5 years] vs 11.0 years [5.5 to 15.5 years]; P � .05).lthough there was no difference between the 2 groups ineight standard deviation score (SDS) at the start of rhGHreatment (median, �2.43 years [�3.8 to 1.5 years] vs �2.44ears [�4.1 to 0.3 years]), height velocity before treatmentas greater in the IGHD group (median, 4.43 cm/year [2.6 to.6 cm/year] vs 3.80 cm/year [1.7 to 5.4 cm/year]; P � .001).eak growth hormone levels on GH stimulation testing alsoere greater in the IGHD group (median, 5.95 mU/L [1.5 to.1 mU/L] vs 3.55 mU/L [0.5 to 10.1 mU/L]; P � .001). Theubjects in the TBI group had a greater median body massndex (BMI) SDS before treatment (�0.26 [�1.6 to 1.6] vs.23 [�1.6 to 1.9]; P � .001).

Data regarding the presence of other pituitary hormoneeficiencies are given in Table II. A statistically significantlyreater frequency of multiple (2 or more) pituitary deficienciesas found in the TBI group. In the TBI group, thyroid-

timulating hormone (TSH) deficiency was the second-mostommonly diagnosed deficiency (30.5%) after GHD, fol-owed by adrenocorticotropic hormone deficiency (23.9%),onadotrophin deficiency (16.9%), and antidiuretic hormoneeficiency (9.3%).

The response to rhGH therapy during the first year ofreatment in the 2 groups also was compared (Table III). TheGHD group received a slightly larger (statistically signifi-antly so) median rhGH dose (0.19 mg/kg/week) than theBI group (0.18 mg/kg/week; P � .016). Although theedian height velocity during this year of therapy was higher

n the IGHD group (5.62 cm/year [1.4 to 9.2 cm/year] vs.59 cm/year [1.0 to 9.8 cm/year]; P � .007), the response to

TBIh to 90th

rcentile n Median10th to 90th

percentile P

4 to 14.5 141 11.0 5.5 to 15.5 .045141 (87.4%) NS

8 to �1.5 141 �2.44 �4.1 to �0.3 NS6 to 6.6 51 3.80 1.7 to 5.4 �.0015 to 9.1 130 3.55 0.5 to 10.1 �.0016 to 1.6 141 0.23 �1.6 to 1.9 �.001

TBI

P% n %

87.4% 134 58.9% �.0018.9% 134 36.2% �.0017.4% 130 30.5% �.0013.1% 129 23.9% �.0013.2% 121 16.9% �.0010.7% 125 9.3% �.001

tics

10t

pe

4.

�3.2.1.

�1.

HD

reatment, as measured by median change in height velocity

Deficiency 591

(Tcsa

aricseowm

cCtrpop

tdfbtcdTbittatfsc

oicfdwa

Tpcatneiwot1dgkcgTtephDaapaot

shts

T

TGHC

B

N

5

�ht vel) over the first year of treatment, was greater in theBI group (3.72 cm/year [0.82 to 7.42 cm/year] vs 4.87

m/year [1.5 to 8.17 cm/year]; P � .016). There were noignificant differences in BMI SDS between the 2 groupsfter the first year of treatment.

DISCUSSIONAlthough data from the KIGS database are selective

nd may not be representative of all children and adolescentseceiving rhGH therapy for GHD, it gives valuable insightnto the current epidemiology of TBI-induced GHD and thelinical characteristics of young patients who present with thispecific problem. Our analysis of the database does not, how-ver, provide us with any information regarding the incidencer prevalence of TBI-induced GHD, nor does it provide usith data on post-TBI survivors who do not have GHD butay have other hypothalamic-pituitary dysfunction.

The most striking observation is the small number ofhildren and adolescents registered with TBI-induced GHD.onsidering the epidemiology of TBI in this age group, and

he frequency of GH deficiency (23% to 69%) reported fromecent studies in adult TBI survivors,2-8 we would have ex-ected many more cases. In the UK, the reported prevalencef severe TBI in children up to age 14 years is 5.6 per 100,000er year, with a survival rate of 76%.12

Information from any large clinical-epidemiologic da-abase has potential limitations with respect to the accuracy ofata provided by clinicians. Moreover, KIGS involved noormal standardization of diagnostic tests; thus, any directias of case registration cannot be excluded. Nevertheless,hese factors are unlikely to account for the small number ofases registered as TBI-induced GHD. It is possible that ourata may truly reflect an underlying low prevalence of post–BI-induced GHD in children and adolescents. This woulde in contrast to the data reported in adults after headnjury,2-8 and would suggest that in children and adolescents,he hypothalamic-pituitary structures may be less susceptibleo damage or may have better capacity for repair and recoveryfter moderate to severe head injury. Evidence in support ofhis hypothesis is lacking, however. Postmortem studies per-ormed immediately after death due to TBI have found equaltructural damage to the hypothalamic-pituitary structures in

able III. Response to rhGH therapy during the first

n M

ime on GH treatment (years) 23,722H dose over first year of treatment (mg/kg/week) 18,529eight velocity (cm/year) 17,903hange in height velocity over first year oftreatment (cm/year)

7400

MI (SDS) at end of 1st year of treatment 23,477

S, not significant.

hild and adult TBI victims.13 Furthermore, there is no w

92 McDonald et al

bjective evidence indicating that in children, either the brainn general or the hypothalamic-pituitary structures specificallyonfer any advantage in terms of neuroplasticity and abilityor brain repair after head injury compared with adults. In-eed, there actually is some evidence suggesting that childrenho suffer TBI at a younger age have less capacity for repair

nd regeneration.14-17

Another possible explanation for the low prevalence ofBI-induced GHD is associated with the much longer delayhase between the head injury event and the development oflinically evident pituitary dysfunction seen in children anddolescents. Prospective longitudinal data in adults suggesthat abnormalities in pituitary function after head injury mayot become evident until at least 6 to 12 months after the TBIvent.2-8 In addition, case reviews of TBI-induced hypopitu-tarism in both adults and children have shown that manyent undiagnosed for 5 years or more.18,19 The median age ofnset of rhGH treatment in KIGS was only slightly higher inhe TBI group compared with the IGHD group (10.3 years vs1.0 years). The median age of occurrence of severe TBIuring childhood is 8.8 years,12 which implies that the TBIroup may have experienced delays in diagnosis of GHD. Ineeping with this, pretreatment height velocity was signifi-antly lower and BMI SDS significantly higher in the TBIroup compared with the IGHD group, consistent with theBI group’s relatively longer duration of GHD status before

he initiation of rhGH therapy. Nevertheless, we cannotxclude the possibility that the increased prevalence of otherituitary hormone deficiencies in the TBI group also mayave adversely affected growth before initiation of treatment.eficiencies in more than 1 hypothalamic-pituitary hormone

xis were observed more frequently in the TBI group.20 Indults, GHD appears to be the most common abnormalityost-TBI, followed by deficiencies in the gonadotrophin,drenal, and, finally, thyroid axes.20 This is comparable withur findings, except for that in our study, TSH deficiency washe second most common deficiency after GHD.

It also is entirely possible that many childhood TBIurvivors, particularly those left with a significant physicalandicap, are being overlooked with respect to assessment ofheir growth and GH status in the post-TBI rehabilitationetting. Children with GHD may have height measurements

r of therapy

HD TBI

n10th to 90th

percentile n Median10th to 90th

percentile P

0.6 to 8.1 141 3.84 0.9 to 8.6 .0149 0.13 to 0.31 138 0.18 0.11 to 0.27 .0162 1.4 to 9.2 108 4.59 1.0 to 9.18 .0072 0.82 to 7.41 39 4.87 1.5 to 8.17 .016

1 �1.67 to 1.59 138 0.14 �1.62 to 1.96 NS

yea

IG

edia

3.10.15.63.7

�0.1

ithin the normal population centile range but with an un-

The Journal of Pediatrics • April 2008

dwsFmtw

rcapamigh

electfishsbdtdobbhT

12ip

3nE4p25et6ea7et8h39pp1f1m21t21p11m1s21h11eB1h11dl2i2

T

erlying growth rate that is inappropriately low and falteringith time and will go undetected unless serial height mea-

urements are made and assessed at appropriate intervals.urthermore, a subset of post-TBI children may exist whoaintain a normal growth rate despite inadequate GH secre-

ion, a phenomenon that has been described in some childrenith GHD, particularly after cranial irradiation.21

The most likely explanation for the low number ofegistered cases of TBI-induced GHD in our study may belinicians’ poor awareness of and failure to recognize TBI as

cause of GHD. Direct questions regarding a history ofrevious head injury may be overlooked by clinicians in theirssessment of children found to have GHD, who are thenistakenly assigned to another GHD diagnostic category. It

s noteworthy that in the KIGS database, 12% of the IGHDroup also exhibited evidence of dysfunction of at least 1 otherypothalamic-pituitary hormone axis.

Although this observation may reflect differences in thetiology and natural history of post–TBI-induced hypotha-amic-pituitary dysfunction between adults and children, it isntirely plausible that TBI is being neglected as a potentialause of GHD in childhood and that evaluation for hypopi-uitarism is being overlooked in TBI survivors. Thus, ourndings underscore the need for prospective longitudinaltudies in children to establish the frequency and naturalistory of endocrine defects occurring after TBI. In addition,tudies addressing the relationship between mechanism ofrain injury (primary or secondary) and pituitary hormoneisturbance are needed. In the meantime, given the poten-ially significant impact of undiagnosed partial or completeeficiency in any of the pituitary hormone systems (not onlyn growth and pubertal development, but also in regard torain growth and neurocognitive function), clinicians need toe made increasingly aware of TBI as a possible cause ofypopituitarism in both the pediatric endocrinology and post-BI rehabilitation settings.

REFERENCES. Escamilla RF, Lisser H. Simmonds disease. J Clin Endocrinol 1942;2:65-96.. Kelly DF, Gonzalo IT, Cohan P, Berman N, Swerdloff R, Wang C. Hypopitu-tarism following traumatic brain injury and aneurysmal subarachnoid hemorrhage: a

reliminary report. J Neurosurg 2000;93:743-52. o

raumatic Brain Injury is a Rarely Reported Cause of Growth Hormone

. Lieberman SA, Oberoi AL, Gilkison CR, Masel BE, Urban RJ. Prevalence ofeuroendocrine dysfunction in patients recovering from traumatic brain injury. J Clinndocrinol Metab 2001;86:2752-6.. Agha A, Rogers B, Sherlock M, O’Kelly P, Tormey W, Phillips J, et al. Anteriorituitary dysfunction in survivors of traumatic brain injury. J Clin Endocrinol Metab004;89:4929-36.. Bondanelli M, De Marinis L, Ambrosio MR, Monesi M, Valle D, Zatelli MC,t al. Occurrence of pituitary dysfunction following traumatic brain injury. J Neuro-rauma 2004;21:685-96.. Leal-Cerro A, Flores JM, Rincon M, Murillo F, Pujol M, Garcia-Pesquera F,t al. Prevalence of hypopituitarism and growth hormone deficiency in adults long-termfter severe traumatic brain injury. Clin Endocrinol (Oxf) 2005;62:525-32.. Aimaretti G, Ambrosio MR, Di Somma C, Gasperi M, Cannavo S, Scaroni C,t al. Residual pituitary function after brain injury–induced hypopituitarism: a prospec-ive 12-month study. J Clin Endocrinol Metab 2005;90:6085-92.. Agha A, Sherlock M, Phillips J, Tormey W, Thompson CJ. The naturalistory of posttraumatic neurohypophysial dysfunction. Eur J Endocrinol 2005;152:71-7.. Acerini CL, Tasker RC, Bellone S, Bona G, Thompson CJ, Savage MO. Hy-opituitarism in childhood and adolescence following traumatic brain injury: the case forrospective endocrine investigation. Eur J Endocrinol 2006;155:663-9.0. Einaudi S, Bondone C. The effects of head trauma on hypothalamic-pituitaryunction in children and adolescents. Curr Opin Pediatr 2007;19:465-70.1. Niederland T, Makovi H, Gal V, Andreka B, Abraham CS, Kovacs J. Abnor-alities of pituitary function after traumatic brain injury in children. J Neurotrauma

007;24:119-27.2. Parslow RC, Morris KP, Tasker RC, Forsyth RJ, Hawley CA. Epidemiology ofraumatic brain injury in children receiving intensive care in the UK. Arch Dis Child005;90:1182-7.3. Harper CG, Doyle D, Adams JH, Graham DI. Analysis of abnormalities inituitary gland in non-missile head injury: study of 100 consecutive cases. J Clin Pathol986;39:769-73.4. Anderson V, Catroppa C. Recovery of executive skills following paediatric trau-atic brain injury (TBI): a 2-year follow-up. Brain Inj 2005;19:459-70.

5. Lehnung M, Leplow B, Herzog A, Benz B, Ritz A, Stolze H, et al. Children’spatial behavior is differentially affected after traumatic brain injury. Child Neuropsychol001;7:59-71.6. Taylor HG, Alden J. Age-related differences in outcomes following child-ood brain insults: an introduction and overview. J Int Neuropsychol Soc997;3:555-67.7. Verger K, Junque C, Jurado MA, Tresserras P, Bartumeus F, Nogues P, et al. Ageffects on long-term neuropsychological outcome in paediatric traumatic brain injury.rain Inj 2000;14:495-503.8. Benvenga S, Campenni A, Ruggeri RM, Trimarchi F. Clinical review 113:ypopituitarism secondary to head trauma. J Clin Endocrinol Metab 2000;85:353-61.9. Dogru O, Koken R, Bukulmez A, Melek H, Ovali F, Albayrak R. Delay iniagnosis of hypopituitarism after traumatic head injury: a case report and review of the

iterature. Neuro Endocrinol Lett 2005;26:311-3.0. Agha A, Thompson CJ. Anterior pituitary dysfunction following traumatic brainnjury (TBI). Clin Endocrinol (Oxf) 2006;64:481-8.1. Rappaport M, Herrero-Backe C, Rappaport ML, Winterfield KM. Head injury

utcome up to ten years later. Arch Phys Med Rehabil 1989;70:885-92.

Deficiency 593