Pituitary Dysfunction after Blast Traumatic Brain Injury: The UK BIOSAP Study

ORIGINAL ARTICLE

Pituitary Dysfunction after BlastTraumatic Brain Injury: The UK

BIOSAP Study

David Baxter, MD,1,2 David J. Sharp, MD, PhD,1 Claire Feeney, MD,1,3

Debbie Papadopoulou, BSc, RN,3 Timothy E. Ham, MD,1 Sagar Jilka, BSc, MRes,1

Peter J. Hellyer, BSc, MRes,1 Maneesh C. Patel, BSc, MD,4

Alexander N. Bennett, MD, PhD,5 Alan Mistlin, MD,5 Emer McGilloway, MD,5

Mark Midwinter, MD,2,6 and Anthony P. Goldstone, MD, PhD3,7

Objective: Pituitary dysfunction is a recognized consequence of traumatic brain injury (TBI) that causes cognitive,psychological, and metabolic impairment. Hormone replacement offers a therapeutic opportunity. Blast TBI (bTBI)from improvised explosive devices is commonly seen in soldiers returning from recent conflicts. We investigated: (1)the prevalence and consequences of pituitary dysfunction following moderate to severe bTBI and (2) whether it isassociated with particular patterns of brain injury.Methods: Nineteen male soldiers with moderate to severe bTBI (median age 5 28.3 years) and 39 male controlswith moderate to severe nonblast TBI (nbTBI; median age 5 32.3 years) underwent full dynamic endocrine assess-ment between 2 and 48 months after injury. In addition, soldiers had structural brain magnetic resonance imaging,including diffusion tensor imaging (DTI), and cognitive assessment.Results: Six of 19 (32.0%) soldiers with bTBI, but only 1 of 39 (2.6%) nbTBI controls, had anterior pituitary dysfunction(p 5 0.004). Two soldiers had hyperprolactinemia, 2 had growth hormone (GH) deficiency, 1 had adrenocorticotropichormone (ACTH) deficiency, and 1 had combined GH=ACTH=gonadotrophin deficiency. DTI measures of white mat-ter structure showed greater traumatic axonal injury in the cerebellum and corpus callosum in those soldiers withpituitary dysfunction than in those without. Soldiers with pituitary dysfunction after bTBI also had a higher prevalenceof skull=facial fractures and worse cognitive function. Four soldiers (21.1%) commenced hormone replacement(s) forhypopituitarism.Interpretation: We reveal a high prevalence of anterior pituitary dysfunction in soldiers suffering moderate to severebTBI, which was more frequent than in a matched group of civilian moderate to severe nbTBI subjects. We recom-mend that all patients with moderate to severe bTBI should routinely have comprehensive assessment of endocrinefunction.

ANN NEUROL 2013;00:000–000

The use of improvised explosive devices (IEDs) has

characterized the Iraq and Afghanistan conflicts, with

blast traumatic brain injury (bTBI) a “signature injury.”1

More than 400 UK and 2,000 US soldiers have been

fatally wounded by blast injuries in Afghanistan since

2001.2 Among survivors, it is estimated that 19.5% of

1.64 million US troops deployed in both conflicts have

suffered a probable bTBI.3 Soldiers are usually young, so

View this article online at wileyonlinelibrary.com. DOI: 10.1002/ana.23958

Received Mar 8, 2013, and in revised form May 8, 2013. Accepted for publication May 24, 2013.

Address correspondence to Dr Goldstone, Metabolic and Molecular Imaging Group, MRC Clinical Sciences Centre, Imperial College London, Hammer-

smith Hospital, Du Cane Road, London W12 0NN, United Kingdom. E-mail: [email protected]

From the 1Computational, Cognitive, and Clinical Neuroimaging Laboratory, Division of Brain Sciences, Imperial College London, Hammersmith Hospi-

tal, London; 2Royal Centre for Defence Medicine, Academic Department of Military Surgery and Trauma, Birmingham; 3Imperial Centre for Endocrinol-

ogy, Imperial College Healthcare NHS Trust, Charing Cross Hospital, London; 4Imaging Department, Imperial College Healthcare NHS Trust, Charing

Cross Hospital, London; 5Defence Medical Rehabilitation Centre, Headley Court, Epsom, Surrey; 6Academic Section for Musculoskeletal Disease,

Chapel Allerton Hospital, University of Leeds, Leeds; and 7Metabolic and Molecular Imaging Group, Medical Research Council Clinical Sciences Centre,

Imperial College London, Hammersmith Hospital, London, United Kingdom.

Additional supporting information can be found in the online version of this article.

VC 2013 American Neurological Association 1

the long-term impact of consequent physical, cognitive,

and psychological problems represents a significant health

burden. There are no current pharmaceutical treatments

that improve recovery following TBI.4

Nonblast TBI (nbTBI) is a recognized cause of

pituitary dysfunction, in particular growth hormone

(GH) deficiency.5 Reported prevalence rates of pituitary

dysfunction following nbTBI vary between 2 and

68%.5,6 This variability is due in part to differences in

the normal ranges and dynamic endocrine tests used, the

time since injury, and injury severity.5–7 In addition to

adverse metabolic consequences, hypopituitarism causes

multiple symptoms impacting on physical and psycholog-

ical well-being that will impair recovery after TBI, and

thus hormone replacement represents an important thera-

peutic opportunity.8–11 It is unknown how often bTBI

leads to pituitary dysfunction.12

Diffusion tensor imaging (DTI) is a sensitive mag-

netic resonance (MR) technique that can assess the pres-

ence and severity of white matter damage after TBI.13,14

TBI alters the pattern of water diffusion within white

matter, resulting in abnormal diffusion measures, includ-

ing fractional anisotropy (FA). DTI abnormalities in sev-

eral brain regions have been reported in soldiers

following mild bTBI.15 We hypothesized that DTI

would reveal differences in white matter damage in those

soldiers with pituitary dysfunction after bTBI.

Here we report findings from the UK BIOSAP

(United Kingdom Blast Injury Outcome Study of Armed

Forces Personnel). We investigated the prevalence and

associations of pituitary dysfunction in soldiers after

moderate to severe bTBI compared to a control group of

patients after nbTBI.

Subjects and Methods

RecruitmentNineteen military bTBI patients were recruited using the

Academic Department of Military Emergency Medicine

(Birmingham, UK) trauma database to identify soldiers

injured between December 2009 and March 2012. This rep-

resents 10.4% of the 183 UK soldiers who had survived a

moderate to severe bTBI in Afghanistan during this 27-

month period, of what is now the 12th year of this conflict.

Research ethics committee approval and informed consent

were obtained.

Comparison was made with an age- and gender-matched

control group of 39 patients after nbTBI. This represented all

the patients seen in our multidisciplinary Traumatic Brain

Injury clinic at Charing Cross Hospital, London, United King-

dom between August 2009 and March 2012 who met all inclu-

sion=exclusion criteria and were within the age range of the

bTBI group. These patients had identical endocrine assessment

as part of their routine clinical care.

The inclusion criterion for bTBI was a moderate to

severe brain injury caused directly by a single exposure to a

blast. To better examine the effects of the primary blast wave

only, exclusion criteria for bTBI were: (1) requirement for mas-

sive blood transfusion; (2) intracranial lesions causing mass

effect; and (3) post-traumatic stress disorder (PTSD), because

this has been linked with endocrine disturbance.16,17 PTSD was

diagnosed on the basis of psychologist interview and, if sus-

pected, subsequent self-reported symptom ratings from the

PTSD Checklist–Military version derived from Diagnostic and

Statistical Manual of Mental Disorders, 4th edition criteria.18

Although this includes symptoms present in many soldiers after

bTBI, such as loss of memory of the event, anhedonia, social

isolation, sleep disturbance, emotional lability, and poor con-

centration, subjects did not display additional symptoms

required for the diagnosis of PTSD, such as “repeated, disturb-

ing memories, thoughts, images or dreams of a previous stress-

ful experience” or “physical reactions (such as heart pounding,

trouble breathing or sweating) when reminded of a previous

stressful experience.”

Inclusion criteria for both bTBI and nbTBI were: (1)

male gender, (2) >2 and <48 months from a single TBI, (3)

moderate to severe brain injury using the Mayo classification

criteria,19 (4) ongoing cognitive and=or psychological symp-

toms, and (5) completion of all endocrine testing. Exclusion

criteria for bTBI and nbTBI subjects were: (1) diabetes melli-

tus, (2) pre-TBI history of psychiatric disorder, (3) current or

previous drug or excess alcohol use, (4) reversed sleep–wake

cycle, and (4) craniotomy following injury (to avoid the diffi-

culties in brain image registration resulting from gross changes

in brain structure).

Both bTBI and nbTBI subjects underwent clinical assess-

ment and calculation of Abbreviated Injury Score (AIS) and

total Injury Severity Score (ISS), and completed quality of life

(QoL) and symptom questionnaires (see Supplementary

Methods).

Endocrine TestingThe algorithm used to define pituitary dysfunction is shown in

Table 1 (see Supplementary Methods). All patients had mea-

surement of basal serum anterior pituitary hormones followed

by dynamic endocrine testing. Initial screening for GH and

adrenocorticotropic hormone (ACTH) deficiency used the glu-

cagon stimulation test (GST).20,21 The diagnosis of GH defi-

ciency was confirmed with second-line growth hormone-

releasing hormone (GHRH)–arginine test and=or insulin toler-

ance test (ITT).10,22,23 ACTH deficiency was confirmed with

an ITT or metyrapone stimulation test, together with a cortisol

day curve.21,24 Symptoms of diabetes insipidus were investi-

gated further with a water deprivation test.

Cognitive Function AssessmentEach soldier with bTBI completed a standardized neuropsycho-

logical test battery previously shown to be sensitive to cognitive

impairment after TBI.14 The tests looked at the cognitive

domains of: (1) current verbal and nonverbal reasoning ability;

ANNALS of Neurology

2 Volume 00, No. 00

(2) associative memory and learning; (3) executive functions of

set shifting, inhibitory control, cognitive flexibility, and word

generation fluency; and (4) information processing speed (see

Supplementary Methods).

Structural Brain ImagingEach soldier had standard T1, gradient-echo (T2*), and

susceptibility-weighted MR imaging (MRI) to assess focal brain

injury, microbleeds, superficial siderosis, gliosis, contusions, and

DTI. Most patients with pituitary dysfunction also had a pitui-

tary MRI with gadolinium contrast to look for more detailed

hypothalamic–pituitary abnormalities. Patients with nbTBI had

only computed tomography (CT) brain and=or standard

T1=T2 brain MRI as part of routine clinical practice. DTI

analysis of white matter tracts combined tract-based spatial sta-

tistics and region of interest (ROI) approaches (Functional

Magnetic Resonance Imaging of the Brain Software Library,

Oxford, UK), focusing on regions previously shown to be sensi-

tive to damage in bTBI and nbTBI (Supplementary Fig S1 and

Supplementary Methods).14,15 This allowed assessment of

regional FA, a measure of traumatic axonal injury.

Statistical AnalysesComparisons between groups (nbTBI vs bTBI; and bTBI with

pituitary dysfunction vs bTBI without pituitary dysfunction)

were made using Fisher exact test for prevalence data, and

unpaired Student t test (FA and neurocognitive variables), or

Mann–Whitney U test (other variables) for continuous data

(SPSS v19.0; IBM, Armonk, NY). Significance was defined as p

< 0.05. A group 3 ROI repeated measure analysis of variance

was performed to assess the overall effect of pituitary dysfunc-

tion on FA.

Results

Patient CharacteristicsAll soldiers with bTBI had been injured by IEDs and

had been wearing full personal protective equipment. All

required immediate transfer to Camp Bastion for emer-

gency medical treatment, and repatriation to the United

Kingdom within 48 hours. We have detailed information

about the blast exposure, but for operational security rea-

sons these cannot be reported. In the control nbTBI

TABLE 1. Diagnostic Algorithm for Pituitary Dysfunction

Pituitary Axis First Test Confirmatory Test

GH deficiency Glucagon stimulation test: peak GH <5lg/l

GHRH–arginine test: GH < cutoffbased on age and BMI;22 OR ITT:peak GH < 3lg/l

ACTH deficiency Glucagon stimulation test: peak corti-sol < 350nmol/l (<12.7lg/dl)21

Metyrapone test: 11-DOC <200nmol/l (<6.9lg/dl) OR if unavail-able ACTH < 60ng/l despite cortisol< 200nmol/l (<7.2lg/dl); OR ITT:peak cortisol < 450nmol/l (<16.3lg/dl); supported by AM cortisol <100nmol/l (<3.62lg/dl)

Hyperprolactinemia Prolactin > 375 mU/l (NR 5 75–375)

Repeat prolactin > 375mU/l ANDnegative macroprolactin AND normalMRI pituitary with contrast

Gonadotrophin deficiency Random testosterone < 10nmol/l(<2.9ng/ml) OR if SHBG low(<15nmol/l) FAI < 30; AND nonele-vated LH (NR 5 1.7–12.0 IU/l) andFSH (NR 5 1.7–8.0 IU/l)

Repeat abnormal basal levels usingmorning (9–10 AM) sample

TSH deficiency Free T4 < 9.0pmol/l (<0.70ng/dl)OR free T3 < 2.5pmol/l (<0.16ng/dl); AND nonelevated TSH (NR 50.30–4.22mU/l)

Repeat abnormal basal levels

ADH (vasopressin) deficiency(diabetes insipidus)

Symptoms of polyuria or polydipsiaAND random urine osmolarity <750mosmol/kg

Water deprivation test

11-DOC 5 11-deoxycorticosterone; ACTH 5 adrenocorticotropic hormone; ADH 5 antidiuretic hormone; BMI 5 body massindex; FAI 5 free androgen index (100 3 testosterone/SHBG); FSH 5 follicle-stimulating hormone; GH 5 growth hormone;GHRH 5 growth hormone-releasing hormone; ITT 5 insulin tolerance test; LH 5 luteinizing hormone; MRI 5 magnetic reso-nance imaging; NR 5 normal range; SHBG 5 sex hormone-binding globulin; TSH 5 thyroid-stimulating hormone.

Baxter et al: Pituitary Dysfunction and TBI

Month 2013 3

TABLE 2. Patient Characteristics

Characteristic MaximumScore

All nbTBI All bTBI p bTBI: NoPituitaryDysfunction

bTBI:PituitaryDysfunction

p

No. 39 19 13 6

Age at TBI, yr 31.3 [22.5–35.7] 26.7 [26.1–30.9] 0.40 26.6 [24.6–30.6] 29.3 [25.8–36.6] 0.48

17.2–44.8 19.0–43.5 19.0–36.3 25.0–43.5

Age at testing, yr 32.3 [23.1–36.7] 28.3 [26.8–32.2] 0.40 28.0 [25.3–31.4] 30.3 [27.4–38.3] 0.32

19.9–45.1 19.6–44.7 19.6–37.6 26.3–44.7

Time sinceTBI, mo

5.8 [3.1–11.0] 15.2 [10.8–19.3] 0.001a 15.2 [8.8–16.6] 17.6 [12.3–20.2] 0.32

1.9–41.2 4.1–23.7 4.1–23.7 4.9–21.9

ISS 75 25.0 [16.0–32.0]1–75

33.0 [20.0–45.0]9–70

0.17 24.0 [14.5–40.5]9–45

35.5 [27.0–51.3]9–70

0.24

AIS head 6 5.0 [4.0–5.0]1–6

4.0 [3.0–5.0]0–6

0.04a 4.0 [2.5–4.0]0–5

5.0 [3.0–5.3]0–6

0.06

AIS chest 6 0 [0–0]0–6

0 [0–2]0–4

0.11 0 [0–3]0–4

0.5 [0–2.3]0–3

0.83

AIS abdomen 6 0 [0–0]0–3

0 [0–2]0–3

0.02a 0 [0–2]0–2

0 [0–2.3]0–3

0.97

GCS 15 14.0 [6.0–14.0]b

3–153.0 [3.0–14.5]c

3–150.24 14.0 [3.0–15.0]d

3–153.0 [3.0–3.0]e

3–30.19

PTA, days 0.5 [0–7.3]f

0–425.5 [0.8–22.8]0–84

0.01a 3.0 [0–19.3]0–84

15.5 [6.3–31.5]4–42

0.10

PTA > 24 hours 20 (51.3%) 13 (68.4%) 0.27 7 (58.3%) 6 (100%) 0.11

BMI, kg/m2 24.7 [22.4–29.4]17.0–33.4

26.7 [24.5–28.9]21.7–33.7

0.28 26.6 [24.5–28.7]g

23.6–29.425.5 [22.4–32.0]h

21.7–33.70.79

Limb amputation 0 (0%) 8 (42.1%) <0.001a 6 (46.1%) 2 (33.3%) 1.00

Major organdamage

3 (7.7%) 11 (57.9%) <0.001a 7 (53.9%) 4 (66.7%) 1.00

Skull/facial fracture 6 (15.4%) 3 (15.8%) 1.00 0 (0%) 3 (50.0%) 0.02

Opiate use 3 (7.7%) 9 (47.3%) 0.001a 6 (46.2%) 3 (50.0%) 1.00

Antidepressant use 5 (12.8%)i 10 (52.7%)j 0.003a 7 (53.8%)k 3 (50.0%)l 1.00

Seizures post-TBI 3 (7.7%)m 2 (10.5%)n 1.00 1 (7.7%)o 1 (16.7%)p 1.00

Primaryhypogonadism

1 (2.6%)q 4 (21.1%)r 0.04a 4 (30.8%)r 0 (0%)r 0.26

Data are expressed as median [interquartile range], range, or No. (%). Probability values are from Mann–Whitney U test or Fisherexact test between groups.aStatistically significant; p < 0.05.Data available for bn 5 16, cn 5 9, dn 5 5, en 5 4, fn 5 38, and due to amputations: gn 5 7, hn 5 4.For analgesic purposes only in: in 5 5 (12.8%), jn 5 6 (31.6%), kn 5 4 (30.8%), ln 5 2 (33.3%).For depression itself in: in 5 0 (0%), jn 5 4 (21.1%), kn 5 3 (23.1%), ln 5 1 (16.7%).On antiepileptic drugs in mn 5 3, nn 5 1, on 5 0, pn 5 1.qNot due to trauma.rDue to perineal trauma.AIS 5 Abbreviated Injury Score; BMI 5 body mass index; bTBI 5 blast traumatic brain injury; GCS 5 Glasgow Coma Scale;ISS 5 Injury Severity Score; nbTBI 5 nonblast TBI; PTA 5 post-traumatic amnesia.

ANNALS of Neurology

4 Volume 00, No. 00

group, injuries were secondary to road traffic accidents

(RTAs; 43%), assaults (32%), falls (23%), and sporting

injuries (2%). Three subjects in the nbTBI group had

experienced previous TBI (1 subject had 2 mild TBIs

from an RTA and an assault, 1 a mild TBI from a fall,

and 1 a TBI of unknown severity from an assault).

The bTBI and nbTBI groups were well matched in

most respects (Table 2). There were no significant differ-

ences in age, ISS whole body injury severity, skull=facial

fractures (15.8 vs 15.4%), or post-traumatic seizures

(10.5 vs 7.7%). The bTBI group had longer post-

traumatic amnesia (PTA; median 5.5 days vs 0.5 days, p5 0.01); more injuries requiring surgery to or loss of

function of major extracranial organs (57.9 vs 7.7%, p 5

0.002); more amputations (36.8 vs 0%, p < 0.001); and,

in keeping with this, more use of strong prescription opi-

ates (47.3 vs 7.7%, p 5 0.001). The time from TBI to

endocrine testing was significantly longer in the bTBI

group (median 15.2 vs 5.8 months, p 5 0.001).

Prevalence of Pituitary Function in bTBI andnbTBI CohortsSix of 19 soldiers with bTBI (31.6%) had anterior pitui-

tary dysfunction, compared to only 1 of 39 (2.6%) sub-

jects with nbTBI (p 5 0.004; Fig 1, Supplementary

Tables S1–S3). Two soldiers (10.5%) had monomeric

hyperprolactinemia (without secondary hypogonadism), 1

(5.3%) had isolated ACTH deficiency, 2 (10.5%) had

isolated GH deficiency, and 1 (5.3%) had combined

ACTH, GH, and gonadotrophin deficiencies. The only

pituitary dysfunction noted in 1 patient with nbTBI was

isolated GH deficiency following a single TBI. No

patients in either group had thyroid-stimulating hormone

(TSH) deficiency or diabetes insipidus.

The 3 soldiers with GH deficiency had insulin-like

growth factor-I (IGF-I) levels in the low normal range

(see Supplementary Table S2), and the 2 soldiers with

ACTH deficiency had normal early morning cortisol lev-

els on initial assessment of 287 to 292nmol=l equivalent

to 10.3 to 10.5lg=dl (normal, >150nmol=l, >5.4lg=dl,

respectively; see Supplementary Table S3). However, on

subsequent cortisol day curves, both subjects with ACTH

deficiency had low cortisol levels (<100nmol=l,

3.62lg=dl) at either 9:00 AM or 12:00 PM on a day curve

consistent with the diagnosis (see Supplementary Results,

Supplementary Table S3). Thus, although the less com-

monly used metyrapone test was occasionally performed

as the confirmatory test to diagnose or exclude ACTH

deficiency instead of the gold standard ITT, findings

were always compatible with the results of baseline or

day curve cortisol levels. Furthermore, as with previous

studies, we have found good specificity and concordance

between the results of the metyrapone test compared to

the ITT or ACTH stimulation test for diagnosing

ACTH deficiency (see Supplementary Results). None of

the soldiers with ACTH deficiency had any history of

hypotension, hypoglycemia, or hyponatremia.

Primary hypogonadism due to perineal=testicular

blast injury had been found in an additional 4 of 19 sol-

diers with bTBI (21.2%), none of whom had pituitary dys-

function, and all were already on testosterone replacement

(see Supplementary Results, Supplementary Table S1).

Comparison of bTBI with versus bTBI withoutPituitary DysfunctionThere was no significant difference in age at TBI, time

since injury, ISS, abdominal AIS, body mass index

(BMI), or prevalence of amputations, nonhead major

organ damage, seizures, any use of antidepressants or spe-

cifically for depression, or opiate use between bTBI

patients with versus those without pituitary dysfunction

(see Table 2, Supplementary Tables S6 and S7). BMI

FIGURE 1: Prevalence of pituitary dysfunction in nonblast traumatic brain injury (nbTBI) and blast TBI (bTBI). Greater prevalenceof anterior pituitary dysfunction was seen in subjects after bTBI (right) than nbTBI (left). No subjects had thyroid-stimulatinghormone deficiency or diabetes insipidus. ACTH 5 adrenocorticotropic hormone; GH 5 growth hormone; Gn 5gonadotrophin.


Month 2013 5

could not be adequately assessed in the 8 soldiers with

bTBI who had limb amputations, but none was mor-

bidly obese on clinical examination.

There were trends for the AIS head injury scores to be

higher (p 5 0.06), and duration of PTA to be longer (median

5 15.5 vs 3.0 days, p 5 0.10) in those soldiers with pituitary

dysfunction after bTBI than in those without.

The single soldier (M08) with multiple pituitary

deficiencies was taking opiates at the time of diagnosis of

gonadotrophin deficiency and initial dynamic endocrine

testing with a GST. However, both GH and ACTH defi-

ciency were subsequently confirmed using an ITT after

opiates had been discontinued.

Neuroimaging ResultsIn the bTBI group, we investigated whether particular

structural abnormalities were associated with pituitary

dysfunction. Three of the 6 (50.0%) soldiers with pitui-

tary dysfunction, compared to only 1 of the 13 (7.7%)

soldiers without pituitary dysfunction, had contusions on

brain MRI scans (p 5 0.07). One soldier with pituitary

dysfunction had 2 contusions, whereas the remainder

had 1 contusion (Supplementary Fig S2). The total con-

tusion volume was <10cm3 in all cases; the soldier with-

out pituitary dysfunction had the smallest contusion

volume. There was a greater prevalence of skull=facial

fractures in the soldiers with pituitary dysfunction com-

pared to those without (50 vs 0%, p 5 0.02).

There were no significant differences in the prevalence

of other abnormalities visible on acute CT brain scans fol-

lowing blast exposure or study structural MR scans, includ-

ing presence of extracerebral, subarachnoid, or

intraventricular hemorrhage, microbleeds, superficial sider-

osis, or gliosis, between those soldiers with versus without

pituitary dysfunction (Supplementary Table S4). No hypo-

thalamic–pituitary abnormalities were seen on MRI brain

scans in any soldiers in the bTBI group, or in the 4 with

pituitary dysfunction who had dedicated contrast-enhanced

MRI pituitary scans (M01, M08, M10, M14). This

included all those soldiers with hyperprolactinemia and

multiple pituitary hormone deficiencies.

DTI analysis showed a reduction in FA depending

on the ROI, indicating greater white matter damage, in

those soldiers with pituitary dysfunction after bTBI com-

pared to those without (p 5 0.14 effect of group, p 5

0.02 group 3 ROI interaction). Planned post hoc analy-

sis showed significantly lower FA values for those soldiers

with pituitary dysfunction within the cerebellum (p <

0.05), and body=genu (p < 0.05) and splenium (p 5

0.01) of the corpus callosum (Fig 2).

Symptoms, QoL, and Cognitive FunctionConsistent with their higher prevalence of polytrauma

and amputations, the soldiers with bTBI had significantly

worse scores for physical activity and daily living prob-

lems than the control nbTBI group, but not in measures

of depression and emotional well-being (see Supplemen-

tary Table S5, Supplementary Results).

In the bTBI group, soldiers with pituitary dysfunc-

tion had trends toward worse measures of QoL and

symptom scores in several domains relating to emotional

and social functioning, fatigue, and mood compared to

those without pituitary dysfunction (see Supplementary

Table S5, Supplementary Results).

The bTBI subjects with pituitary dysfunction had

significantly worse average current verbal intellectual abil-

ity than those without pituitary dysfunction, despite

there being no significant difference in their premorbid

intelligence (Wechsler Test of Adult Reading; Table 3).

The bTBI group with pituitary dysfunction also showed

significantly worse cognitive impairment in the domains

of visual=naming=reading processing speed, verbal flu-

ency, and information processing (see Table 3).

Discussion

We have demonstrated a high prevalence of pituitary dys-

function following moderate to severe blast TBI. Almost

a third of soldiers with bTBI had anterior pituitary

abnormalities, compared to only 2% of age- and gender-

matched civilians with moderate to severe nbTBI. The

FIGURE 2: Pituitary dysfunction and white matter damage inblast traumatic brain injury. Lower fractional anisotropy wasseen in a priori white matter tract regions of interest in soldierswith pituitary dysfunction after blast traumatic brain injury(black, n 5 6) compared to those without pituitary dysfunction(white, n 5 13). Data are expressed as mean 6 standard devia-tion. *p < 0.05 (unpaired t test). Ant 5 anterior; CC 5 corpuscallosum; Cap 5 capsule; Int 5 internal; Post 5 posterior; WM5 white matter.

ANNALS of Neurology

6 Volume 00, No. 00

most common pituitary abnormality in bTBI was GH

deficiency, followed by hyperprolactinemia, ACTH, and

gonadotrophin deficiency. One patient had multiple hor-

mone deficiencies.

We carefully avoided overdiagnosis of pituitary dys-

function. We used identical diagnostic algorithms in the

bTBI and nbTBI groups, excluded the presence of mac-

roprolactin, applied strict normal ranges for diagnosing

testosterone and TSH deficiency, performed 2 stimula-

tion tests to confirm ACTH or GH deficiencies, and

adjusted for the confounds of age and obesity in diagnos-

ing GH deficiency.22 This allows us to be confident of

our reported prevalence of pituitary dysfunction in both

groups.6,7

Our results suggest that all patients after moderate

to severe bTBI should undergo endocrine assessment.

Unlike TSH and gonadotrophin deficiency, GH and

ACTH deficiency cannot be excluded or always confirmed

by basal IGF-I or cortisol measurements. Therefore,

dynamic endocrine testing is required. The choice of tests

needs to take into account contraindications for use of the

ITT, such as seizures, as well as the advantages and disad-

vantages of each test, including their specificity=sensitivity,

age=obesity-adjusted normal ranges, resource implications,

local expertise, and drug availability.7,21,23

The presence of pituitary dysfunction after bTBI

was not explicable by differences in age, gender, or obe-

sity. The time to endocrine testing was longer in the

TABLE 3. Pituitary Dysfunction and Cognitive Function in Blast Traumatic Brain Injury

Cognitive Domain Cognitive Variable No PituitaryDysfunction,n 5 13

PituitaryDysfunction,n 5 6

Premorbid intelligence: read-ing ability

WTAR raw score 35.9 6 11.7 34.7 6 14.6

Intellectual ability WASI similarities (verbal) 32.6 6 6.2 27.0 6 4.1a

WASI matrix reasoning(nonverbal)

24.4 6 7.5 24.2 6 6.0

Memory: associative memory People test immediate recall 22.6 6 8.1 25.0 6 7.8

Processing speed: visualsearch/complex

Trail Making Test trail A,seconds

23.1 6 5.7 28.7 6 5.2a

Trail Making Test trail B,seconds

47.9 6 14.5 53.8 6 12.2

Processing speed: naming/reading

Stroop color naming, seconds 32.5 6 9.1 51.0 6 29.7a

Stroop word reading, seconds 24.3 6 6.7 37.2 6 13.6b

Executive function:alternating-switch cost

Trail Making Test trail B 2A, seconds

24.8 6 13.5 25.2 6 9.0

Executive function: cognitiveflexibility

Color word Stroop inhibi-tion/switching, seconds

70.5 6 24.2 86.3 6 30.8

Inhibition/switching minus abaseline of color naming andword reading, seconds

30.0 6 18.8 26.5 6 8.5

Word generation fluency DKEFS letter fluency F 1 A1 S total

40.1 6 12.9 28.8 6 3.6a

Information processing Choice reaction task medianreaction time, milliseconds

413 6 38 473 6 31a

Worse cognitive function was seen in soldiers with pituitary dysfunction after blast traumatic brain injury (n 5 6) compared tothose without pituitary dysfunction (n 5 13). Data are expressed as mean 6 standard deviation. See Supplementary Methods forfurther details on cognitive tests.ap < 0.05,bp < 0.005 (unpaired t test).DKEFS 5 Delis–Kaplan Executive Function System; WASI 5 Wechsler Abbreviated Scale of Intelligence Similarities and MatrixReasoning subsets; WTAR 5 Wechsler Test of Adult Reading.


Month 2013 7

bTBI than nbTBI group. However, this might be

expected to reduce the prevalence of pituitary dysfunc-

tion, as it may resolve over time following TBI.25 Simi-

larly, use of opiates or other medications does not

explain our results. Opiates can have complex neuroen-

docrine effects, including induction of hypogonadotro-

phic hypogonadism, and potentially decreasing ACTH

secretion but increasing GH secretion.26 Although there

was greater use of opiates in the bTBI as a whole than in

the nbTBI group, the individual pituitary dysfunction

seen in each soldier within the bTBI group was not

explicable by opiate use. The bTBI group did have more

polytrauma than the nbTBI group, which may be a con-

tributory factor, although the mechanism linking periph-

eral injury to hypothalamic–pituitary dysfunction is

uncertain.

Blast appears to produce a distinct pattern of

TBI,15,27 although the mechanism by which blast injury

damages the brain remains unclear, limiting our ability

to identify those patients at high risk of pituitary dys-

function. The primary blast wave or wind may cause

direct injury, or secondary injuries from explosion debris

or tertiary injuries from the impact of being thrown by

the blast may occur.28,29 These injuries could affect the

hypothalamus, pituitary gland, or pituitary stalk, result-

ing in damage to cell bodies or white matter connections

as well as hypophyseal vessels, local superficial siderosis,

inflammation, or hypovolemia=ischemia.

Our imaging results do not provide clear evidence

about the precise mechanism of hypothalamic–pituitary

damage. We did not see evidence of focal injury to the

hypothalamus–pituitary or superficial siderosis, and we

excluded bTBI subjects who needed massive blood transfu-

sions. However, pituitary dysfunction may be related to the

severity of brain injury after blast exposure, as suggested in

nbTBI.5 This is supported in our study by the longer dura-

tion of PTA in the bTBI than in the nbTBI group

(although interpretation may be complicated by sedation

and anesthesia), and the presence of more white matter

damage15 and more skull=facial fractures, and a trend for

more cerebral contusions and longer PTA, in those soldiers

with than in those without pituitary dysfunction after bTBI.

Diffuse axonal injury is common in the corpus callosum

after TBI in general,30 and posterior fossa white matter

tracts are particularly damaged after mild bTBI.15 It remains

unclear whether the more severe damage to these tracts in

bTBI with pituitary dysfunction simply indicates a greater

severity of brain injury, or is indicative of a particular injury

pattern associated with hypothalamic–pituitary damage.

Our study focused on subjects with a single epi-

sode of moderate to severe bTBI. It remains to be

determined whether pituitary dysfunction is a significant

problem after single, or especially repeated, mild bTBI,

because there is evidence that multiple bTBI may aug-

ment neurological deficits.31 A single previous study has

suggested that repeated mild bTBI can produce endo-

crine disturbance.32 However, methodological issues

with this study make it difficult to interpret, including

their reliance on basal hormone measurements, the defi-

nition of normal ranges from a small cohort of control

subjects, and the nonstandard assessment of posterior

pituitary function.

The trend for worse fatigue, emotional symptoms,

social problems, and mood in those soldiers with pitui-

tary dysfunction after bTBI may be related to worse

underlying brain injury and=or their endocrine problems.

These are well-recognized features of GH deficiency, and

lethargy is also seen in cortisol and testosterone defi-

ciency.8,9,33 Similarly, cognitive impairment in soldiers

with pituitary dysfunction after bTBI may be related to

both greater brain=axonal injury and hormone deficien-

cies, including GH.14,34,35

Our findings led to substantial changes in clinical

management. The soldier with hypogonadotrophic hypo-

gonadism was treated with intramuscular long-acting tes-

tosterone. Both soldiers with ACTH deficiency were

commenced on hydrocortisone replacement. All 3 soldiers

with GH deficiency were >1 year after bTBI and have

been started on GH replacement in view of persistent neu-

ropsychological symptoms despite replacement of other

pituitary hormones. The soldiers with sufficient follow-up

data available have had a symptomatic improvement after

6 months of GH replacement, with adult growth hor-

mone deficiency QoL assessment (AGHDA-QoL) score

falling from 19 to 14 (of 25), and Beck Depression Inven-

tory II (BDI-II) score from 36 to 18 (of 63) in 1 subject

(M14), and AGHDA-QoL from 14 to 3, and BDI-II fall-

ing from 20 to 16 in another (M08) during this period.

However, the other soldier receiving GH (M07) is still

undergoing dose titration, and so it is too early to assess

his symptomatic improvement. The soldiers with mild

hyperprolactinemia did not require treatment, as second-

ary hypogonadism was absent.

In conclusion, this is the first study to demonstrate

a high prevalence of anterior pituitary hormone abnor-

malities after moderate to severe bTBI. The prevalence

was greater than in a matched group of civilian nbTBI,

suggesting that pituitary dysfunction is a particular prob-

lem after blast exposure. Pituitary dysfunction following

bTBI was associated with worse cognitive function and

greater severity of head injury, including white matter

damage. Given that there were no completely diagnostic

predictors of pituitary dysfunction in bTBI, we recom-

mend that in clinical practice all soldiers with moderate

ANNALS of Neurology

8 Volume 00, No. 00

to severe bTBI undergo routine and comprehensive pitui-

tary function testing during rehabilitation.

Acknowledgment

This study was supported by the UK Medical Research

Council (MRC), National Institute for Health Research

(NIHR, ref. NIHR-RP-011-048), Imperial College Health-

care Charity (ref. 7006/R21U), and the Royal Centre for

Defence Medicine. D.J.S. is supported by the MRC (Clini-

cian Scientist Fellowship) and NIHR, A.P.G. by the MRC,

and C.F. by Imperial College Healthcare Charity.

We thank the trauma nurse coordinators, Academic

Department of Military Surgery and Trauma, Birming-

ham, UK; doctors, nurses, and rehabilitation staff based

at Defence Medical Rehabilitation Centre, Headley

Court, Surrey, UK; E. Hughes and J. Allsop, Robert

Steiner MRI Unit, MRC Clinical Sciences Centre, Ham-

mersmith Hospital, London, UK for assistance with

MRI; endocrinology colleagues at Imperial College

Healthcare NHS Trust, London, and doctors and nursing

staff, Patient Investigation Unit, Charing Cross Hospital,

London and Metabolic Day Ward, St Mary’s Hospital,

London for assistance with endocrine testing; Depart-

ment of Clinical Biochemistry, Imperial College Health-

care NHS Trust, London for performing hormone assays;

and J. Monson for helpful comments. We especially

thank the soldiers from the UK military who after suffer-

ing these life-changing injuries still had the enthusiasm

and spirit to take part in this research.

The opinions expressed are those of the authors and

not the UK Ministry of Defence.

Authorship

Patient recruitment: D.B., D.J.S., T.E.H., A.N.B., A.M.,

E.M., A.P.G.; study design: D.B., D.J.S., M.M., A.P.G.;

data collection: D.B., D.J.S., D.P., T.E.H., S.J., A.P.G.;

data analysis: D.B., D.J.S., C.F., T.E.H., S.J., P.J.H.,

M.C.P., A.P.G.; data interpretation: D.B., D.J.S., C.F.,

A.P.G.; writing the manuscript: D.B., D.J.S., C.F.,

A.P.G.; review and editing the manuscript: all authors.

Potential Conflicts of Interest

D.J.S., A.P.G.: grants=grants pending, Pfizer.

References

1. Benzinger TL, Brody D, Cardin S, et al. Blast-related brain injury:imaging for clinical and research applications: report of the 2008St. Louis workshop. J Neurotrauma 2009;26:2127–2144.

2. Chesser GS. Afghanistan casualties: military forces and civilians.Washington, DC: Congressional Research Service, 2012.

3. Tanielian T, Jaycox HL. Invisible wounds of war: psychological andcognitive injuries, their consequences, and services to assist recov-ery. Arlington, VA: RAND Center for Military Health PolicyResearch, 2008.

4. Ruff RL, Riechers RG. Effective treatment of traumatic brain injury:learning from experience. JAMA 2012;308:2032–2033.

5. Schneider HJ, Kreitschmann-Andermahr I, Ghigo E, et al. Hypo-thalamopituitary dysfunction following traumatic brain injury andaneurysmal subarachnoid hemorrhage: a systematic review. JAMA2007;298:1429–1438.

6. Kokshoorn NE, Smit JW, Nieuwlaat WA, et al. Low prevalence ofhypopituitarism after traumatic brain injury: a multicenter study.Eur J Endocrinol 2011;165:225–231.

7. Kokshoorn NE, Wassenaar MJ, Biermasz NR, et al. Hypopituitar-ism following traumatic brain injury: prevalence is affected by theuse of different dynamic tests and different normal values. Eur JEndocrinol 2010;162:11–18.

8. Salvatori R. Adrenal insufficiency. JAMA 2005;294:2481–2488.

9. Cherrier MM. Testosterone effects on cognition in health and dis-ease. Front Horm Res 2009;37:150–162.

10. Molitch ME, Clemmons DR, Malozowski S, et al. Evaluation andtreatment of adult growth hormone deficiency: an Endocrine Soci-ety clinical practice guideline. J Clin Endocrinol Metab 2011;96:1587–1609.

11. Bondanelli M, Ambrosio MR, Cavazzini L, et al. Anterior pituitaryfunction may predict functional and cognitive outcome in patientswith traumatic brain injury undergoing rehabilitation. J Neuro-trauma 2007;24:1687–1697.

12. Guerrero AF, Alfonso A. Traumatic brain injury-related hypopitui-tarism: a review and recommendations for screening combat vet-erans. Mil Med 2010;175:574–580.

13. MacDonald CL, Dikranian K, Bayly P, et al. Diffusion tensor imagingreliably detects experimental traumatic axonal injury and indicatesapproximate time of injury. J Neurosci 2007;27:11869–11876.

14. Kinnunen KM, Greenwood R, Powell JH, et al. White matter dam-age and cognitive impairment after traumatic brain injury. Brain2011;134:449–463.

15. MacDonald CL, Johnson AM, Cooper D, et al. Detection of blast-related traumatic brain injury in U.S. military personnel. N Engl JMed 2011;364:2091–2100.

16. Pervanidou P, Chrousos GP. Neuroendocrinology of post-traumatic stress disorder. Prog Brain Res 2010;182:149–160.

17. van Liempt S, Vermetten E, Lentjes E, et al. Decreased nocturnalgrowth hormone secretion and sleep fragmentation in combat-related posttraumatic stress disorder; potential predictors ofimpaired memory consolidation. Psychoneuroendocrinology 2011;36:1361–1369.

18. Wilkins KC, Lang AJ, Norman SB. Synthesis of the psychometricproperties of the PTSD checklist (PCL) military, civilian, and spe-cific versions. Depress Anxiety 2011;28:596–606.

19. Malec JF, Brown AW, Leibson CL, et al. The Mayo classificationsystem for traumatic brain injury severity. J Neurotrauma 2007;24:1417–1424.

20. Leong KS, Walker AB, Martin I, et al. An audit of 500 subcutane-ous glucagon stimulation tests to assess growth hormone andACTH secretion in patients with hypothalamic-pituitary disease.Clin Endocrinol (Oxf) 2001;54:463–468.

21. Cegla J, Jones B, Seyani L, et al. Comparison of the overnightmetyrapone and glucagon stimulation tests in the assessment ofsecondary hypoadrenalism. Clin Endocrinol (Oxf) 2013;78:738–742.

22. Colao A, Di SC, Savastano S, et al. A reappraisal of diagnosing GHdeficiency in adults: role of gender, age, waist circumference, andbody mass index. J Clin Endocrinol Metab 2009;94:4414–4422.


Month 2013 9

23. Yuen KC, Biller BM, Molitch ME, et al. Clinical review: Is lack ofrecombinant growth hormone (GH)-releasing hormone in theUnited States a setback or time to consider glucagon testing foradult GH deficiency? J Clin Endocrinol Metab 2009;94:2702–2707.

24. Grossman AB. Clinical review: The diagnosis and management ofcentral hypoadrenalism. J Clin Endocrinol Metab 2010;95:4855–4863.

25. Aimaretti G, Ambrosio MR, Di SC, et al. Residual pituitary functionafter brain injury-induced hypopituitarism: a prospective 12-monthstudy. J Clin Endocrinol Metab 2005;90:6085–6092.

26. Vuong C, Van Uum SH, O’Dell LE, et al. The effects of opioidsand opioid analogs on animal and human endocrine systems.Endocr Rev 2010;31:98–132.

27. Xiong Y, Mahmood A, Chopp M. Animal models of traumaticbrain injury. Nat Rev Neurosci 2013;14:128–142.

28. Cernak I, Noble-Haeusslein LJ. Traumatic brain injury: an overviewof pathobiology with emphasis on military populations. J CerebBlood Flow Metab 2010;30:255–266.

29. Goldstein LE, Fisher AM, Tagge CA, et al. Chronic traumaticencephalopathy in blast-exposed military veterans and a blastneurotrauma mouse model. Sci Transl Med 2012;4:134ra60.

30. Adams JH, Graham DI, Murray LS, et al. Diffuse axonal injury dueto nonmissile head injury in humans: an analysis of 45 cases. AnnNeurol 1982;12:557–563.

31. Ruff RL, Riechers RG, Wang XF, et al. A case-control study exam-ining whether neurological deficits and PTSD in combatveterans are related to episodes of mild TBI. BMJ Open 2012;2:e000312.

32. Wilkinson CW, Pagulayan KF, Petrie EC, et al. High prevalence ofchronic pituitary and target-organ hormone abnormalities afterblast-related mild traumatic brain injury. Front Neurol 2012;3:11.

33. Webb SM. Measurements of quality of life in patients with growthhormone deficiency. J Endocrinol Invest 2008;31:52–55.

34. van Dam PS. Neurocognitive function in adults with growth hor-mone deficiency. Horm Res 2005;64(suppl 3):109–114.

35. Bonnelle V, Leech R, Kinnunen KM, et al. Default mode networkconnectivity predicts sustained attention deficits after traumaticbrain injury. J Neurosci 2011;31:13442–13451.

ANNALS of Neurology

10 Volume 00, No. 00

1

Pituitary Dysfunction after Blast Traumatic Brain Injury: UK BIOSAP Study

David Baxter, David J Sharp, Claire Feeney, Debbie Papadopoulou, Timothy E Ham, Sagar Jilka,

Peter J Hellyer, Maneesh C Patel, Alex Bennett, Alan Mistlin, Emer McGilloway, Mark Midwinter,

Anthony P Goldstone

Supplemental Material Annals of Neurology

SUPPLEMENTAL FIGURES

Figure S1. White matter tract regions of interest

Figure S2. Intra-cerebral contusions following blast traumatic brain injury

SUPPLEMENTAL TABLES

Table S1. Pituitary-gonadal axis, pituitary-thyroid axis and prolactin in blast traumatic brain injury

Table S2. Growth hormone-IGF-I axis in blast traumatic brain injury

Table S3. ACTH-cortisol axis in blast traumatic brain injury

Table S4. Pituitary dysfunction and structural neuroimaging abnormalities in blast traumatic brain

injury

Table S5. Quality of life and symptom questionnaires in non-blast and blast traumatic brain injury

Table S6. Characteristics of soldiers with blast TBI

Table S7. Medications used by soldiers with blast TBI

SUPPLEMENTAL RESULTS

Non-pituitary endocrine diagnoses in bTBI and nbTBI cohorts

IGF-I levels in bTBI patients with GH deficiency

Symptoms, quality of life and cognitive function

Interpretation of metyrapone test

SUPPLEMENTAL METHODS

Recruitment

Endocrine Testing

Glucagon Stimulation Test

GHRH-Arginine Test

Insulin Tolerance Test (ITT)

2

Cortisol Day Curve

Metyrapone Stimulation Test

Water Deprivation Test

Neuropsychological Assessments

Structural Imaging

DTI Analysis

ACKNOWLEDGMENTS

SUPPLEMENTAL REFERENCES

3

SUPPLEMENTAL FIGURES

Figure S1. White matter tract regions of interest

Regions of interest (ROIs) used for determination of fractional anisotropy (FA) in soldiers after blast

traumatic brain injury (bTBI). Individual color masks overlaid onto group average FA map for

soldiers with bTBI (n=19) registered into standard MNI space (using MNI co-ordinates). ROIs are:

(A) anterior internal capsule, (B) posterior internal capsule, (C) cingulum, (D) corpus callosum, (E)

cerebral peduncles, (F) middle cerebellar peduncles, (G) orbitofrontal white matter, (H) uncinate

fasiculi. FA was sampled from areas within a white matter skeleton (not shown) produced by tract

based spatial statistics (TBSS).

4

Figure S2. Intra-cerebral contusions following blast traumatic brain injury

High resolution T1 brain scans (axial sections) in subject space showing contusions (arrows) in

soldiers after blast TBI (A) without pituitary dysfunction, and (B-D) with pituitary dysfunction. Total

contusion volumes for these patients were: (A) 0.2, (B) 9.1, (C) 0.6, (D) 1.0 cm3.

5

SUPPLEMENTAL TABLES

Table S1. Pituitary-gonadal axis, pituitary-thyroid axis and prolactin in blast traumatic brain injury

Abnormal values indicated by grey shading. * on testosterone replacement,

# calculated from 100 x total testosterone/SHBG,

a to convert to μg/mL divide by 3.467,

b to convert to μg/dL divide by 12.87,

c to convert to μg/dL divide by 15.36,

d remained elevated on repeat measurement with negative macroprolactin. P values from

Mann Whitney U test or Fisher’s exact test between groups.

Abbreviations: ACTH: ACTH deficiency, GH: GH deficiency, Gn: Gonadotrophin deficiency, PRL: Hyperprolactinemia

GONADOTROPHIN / TESTOSTERONE AXIS THYROID PROLACTIN

IDSummary of Pituitary

DysfunctionLH FSH Testosterone a SHBG

Free Androgen

Index #Primary

HypogonadismFree T4 b Free T3 c TSH Prolactin

Units IU/L IU/L nmol/L nmol/L pmol/L pmol/L mU/L mU/L

Normal range 2.0-12.0 1.7-8.0 10.0-30.0 15-55 30-150 9.0-26.0 2.5-5.7 0.3-4.2 75-375

No Pituitary Dysfunction n=13

M02 Nil 2.4 1.1 23.7 55.0 43.1 No 15.6 4.4 0.46 97

M04 Nil 4.7 3.9 21.0 33.0 63.6 No 12.2 5.7 2.11 146

M05 Nil 4.4 2.5 15.5 18.0 86.1 No 15.3 3.2 1.78 118

M09 Nil 1.2 4.8 39.5 * 18.0 219.4 Yes 13.4 5.8 1.04 219

M11 Nil 1.6 0.7 12.5 18.0 69.4 No 14.0 4.4 1.43 285

M12 Nil 18.5 36.4 6.2 * 10.0 62.0 Yes 15.5 3.5 1.21 177

M13 Nil 5.8 9.6 13.1 10.0 131.0 No 13.2 5.2 1.33 240

M15 Nil 40.3 60.3 11.6 * 18.0 64.4 Yes 14.0 6.6 0.77 136

M16 Nil 5.5 4.9 22.3 24.0 92.9 No 13.4 4.4 0.96 200

M17 Nil 4.7 3.3 25.2 39.0 64.6 No 16.9 5.2 1.13 131

M18 Nil 0.0 0.1 12.3 * 19.0 64.7 Yes 18.5 4.8 1.25 312

M19 Nil 2.3 1.5 22.0 28.0 78.6 No 15.0 4.7 2.06 183

M20 Nil 3.8 3.6 28.8 32.0 90.0 No 13.2 4.6 0.70 330

Median [IQR] or n (%) 4.4 [2.0-5.6] 3.6 [1.3-7.3] 21.0 [12.4-24.5] 19.0 [18.0-32.5] 69.4 [64.0-91.5] 4 (30.8%) 14.0 [13.3-15.6] 4.7 [4.4-5.5] 1.21 [0.87-1.61] 183 [134-263]

Range 0-40.3 0.1-60.3 6.2-39.5 10.0-55.0 43.1-219.4 12.2-18.5 3.2-6.6 0.46-2.11 97-330

Pituitary dysfunction n=6

M01 PRL 1.8 2.5 21.7 27.0 80.4 No 17.3 4.6 4.18 619 d

M03 ACTH 1.5 1.2 23.8 35.0 68.0 No 16.0 5.5 1.11 126

M07 GH 3.7 2.4 13.1 24.0 54.6 No 14.4 4.9 1.81 172

M08 ACTH/GH/Gn 1.3 1.8 2.0 18.0 11.1 No 15.5 4.3 1.19 199

M10 PRL 2.6 1.3 22.8 26.0 87.7 No 10.8 4.7 2.29 439 d

M14 GH 6.5 3.9 22.4 33.0 67.9 No 12.9 4.0 0.90 216

Median [IQR] or n (%) 2.2 [1.5-4.4] 2.1 [1.3-2.9] 22.1 [10.3-23.1] 26.5 [22.5-33.5] 68.0 [43.7-82.2] 0 (0%) 15.0 [12.4-16.3] 4.7 [4.2-5.1] 1.50 [1.06-2.76] 208 [161-484]

Range 1.3-6.5 1.2-3.9 2.0-23.8 18.0-35.5 11.1-87.7 10.8-17.3 4.0-5.5 0.90-4.18 126-619

P value 0.37 0.28 0.97 0.42 0.47 0.26 0.90 0.70 0.31 0.47

6

Table S2. Growth hormone-IGF-I axis in blast traumatic brain injury

Abnormal values indicated by grey shading. a to convert to ng/mL divide by 0.131, * using age and BMI normal ranges with BMI 25-30 kg/m

2 if not calculable due to

amputation. P values from Mann Whitney U test between groups. Abbreviations: ACTH: ACTH deficiency, BMI: body mass index, GH: GH deficiency, Gn:

Gonadotrophin deficiency, n/a: not applicable, ND: not done, PRL: Hyperprolactinemia.

GROWTH HORMONE / IGF-1 AXIS

Glucagon Stimulation Test GHRH-Arginine Test Insulin Tolerance Test


DysfunctionIGF-I a

IGF-I

age related NR

IGF-I

median of NR

IGF-I

ratio to medianPeak GH IGF-I a BMI GH cut off * Peak GH IGF-I a Peak GH

Units nmol/L nmol/L nmol/L n/a μg/L nmol/L kg/m2μg/L μg/L nmol/L μg/L

Normal range >5 >5


M02 Nil 22.6 14.2-36.9 22.9 0.72 6.55 n/a n/a n/a n/a 23.5 52.10

M04 Nil 58.0 15.2-42.8 25.5 1.84 13.10 n/a n/a n/a n/a n/a n/a


M09 Nil 29.0 18.3-62.8 33.9 0.62 0.98 18.7 n/a 11.7 17.20 n/a n/a

M11 Nil 31.9 16.5-55.1 30.2 1.01 1.56 31.9 26.6 11.7 37.80 n/a n/a

M12 Nil 38.1 15.2-42.8 25.5 1.21 0.64 38.1 n/a 8.1 13.50 n/a n/a

M13 Nil 74.5 15.1-46.5 26.4 2.37 0.69 ND n/a 11.7 17.30 n/a n/a

M15 Nil 27.4 15.2-42.8 25.5 0.87 2.84 30.0 n/a 8.1 23.30 n/a n/a

M16 Nil 31.3 14.2-36.9 22.9 0.99 3.17 29.8 28.7 8.1 27.00 n/a n/a


M18 Nil 17.9 15.2-42.8 25.5 0.57 2.65 19.9 n/a 8.1 10.00 n/a n/a



Median [IQR] 29.0 [19.9-35.0] 0.87 [0.63-1.11] 3.17 [1.27-8.31] 29.9 [19.6-33.5] 17.30 [13.50-27.00]

Range 17.2-74.5 0.55-2.37 0.64-13.10 19.0-38.0 10.0-38.0


M01 PRL 21.6 15.2-42.8 25.5 0.69 6.97 n/a n/a n/a n/a n/a n/a

M03 ACTH 26.7 15.2-42.8 25.5 0.85 0.19 32.1 31.6 8.1 15.40 n/a n/a

M07 GH 23.7 15.0-39.9 24.4 0.75 2.66 26.1 26.1 5.5 3.43 n/a n/a

M08 ACTH/GH/Gn 16.1 13.1-34.7 21.3 0.66 0.08 16.6 n/a 8.1 4.26 ND 0.18

M10 PRL 18.9 15.2-42.8 25.5 0.60 2.78 29.4 24.3 8.1 25.80 n/a n/a

M14 GH 18.2 15.2-42.8 25.5 0.58 0.05 21.8 33.7 5.5 2.70 n/a n/a

Median [IQR] 20.3 [17.7-24.5] 0.68 [0.60-0.78] 1.43 [0.07-3.83] 26.1 [19.2-30.8] 4.26 [3.07-20.6]

Range 16.1-26.7 0.58-0.85 0.05-6.97 16.6-32.1 2.70-25.80

P value 0.07 0.28 0.09 0.54 0.11

7

Table S3. ACTH-cortisol axis in blast traumatic brain injury

Abnormal values indicated by grey shading. To convert to μg/dL: divide

a by 27.59,

b by 28.86. P values from Mann Whitney U test between groups. Abbreviations: 11-

DOC: 11-deoxycortisol, ACTH: ACTH deficiency, GH: GH deficiency, Gn: Gonadotrophin deficiency, n/a: not applicable, ND: not done, PRL: Hyperprolactinemia.

ACTH CORTISOL AXIS

Glucagon Stress Test Day Curve Metyrapone Test (post levels) Insulin Tolerance Test


DysfunctionBasal Cortisol a Peak Cortisol a Basal ACTH Cortisol a Cortisol a ACTH 11-DOC b Basal Cortisol a Peak Cortisol a

Units nmol/L nmol/L ng/L nmol/L nmol/L ng/L nmol/L nmol/L nmol/L

Normal Range 100-500 >350 <50 09, 12, 15, 18, 21h <200 >60 >200 100-500 >500


M02 Nil 230 230 6.0 304, 208, 226, 295, ND n/a n/a n/a 544 636

M04 Nil 283 378 27.4 395, 153, 136, 72, 54 123 247 244.5 n/a n/a

M05 Nil 309 494 39.2 n/a n/a n/a n/a n/a n/a





M15 Nil 376 497 ND n/a n/a n/a n/a n/a n/a






Median [IQR] 323.0 [225.0-392.5] 435.0 [387.5-495.5] 25.8 [17.8-35.7]

Range 192-427 230-626 6.0-92.7


M01 PRL 606 606 164.0 n/a n/a n/a n/a n/a n/a

M03 ACTH 292 292 23.2 67, 50, <20, 28, <20 38 22 87.1 n/a n/a

M07 GH 419 419 24.8 n/a n/a n/a n/a n/a n/a

M08 ACTH/GH/Gn 287 287 18.7 204, 86, 100, 44, 22 n/a n/a n/a 110 268

M10 PRL 109 350 20.6 422, 333, 200, 260, 79 ND 207 200.0 n/a n/a

M14 GH 88 445 32.2 200 at 13h n/a n/a n/a n/a n/a

Median [IQR] 289.5 [103.8-465.8] 384.5 [290.8-485.3] 24.0 [20.1-65.2]

Range 88-606 287-606 18.7-164.0

P value 0.70 0.28 0.75

8

Table S4. Pituitary dysfunction and structural neuroimaging abnormalities in blast traumatic brain injury

Data given as n (%). P values from Fisher’s exact test between groups. Abbreviations: n/a: not applicable, ND: not done.

No pituitary dysfunction Pituitary dysfunction P

n 13 6

Acute CT brain

Extra-dural hemorrhage 0 (0%) 0 (0%) n/a

Sub-dural hemorrhage 0 (0%) 0 (0%) n/a

Traumatic sub-arachnoid or

intra-ventricular hemorrhage 0 (0%) 0 (0%) n/a

Diffuse swelling 1 (7.7%) 0 (0%) 1.00

Study MRI brain

Contusion 1 (7.7%) 3 (50.0%) 0.07

Siderosis 3 (23%) 1 (16.6%) 1.00

Microbleeds 7 (53%) 3 (50%) 1.00

Gliosis 0 (0%) 1 (16.6%) 0.32

Hypo-pituitary damage 0 (0%) 0 (0%) n/a

MRI pituitary with contrast ND 4 normal, 2 ND n/a

9

Table S5. Quality of life and symptom questionnaires in non-blast and blast traumatic brain injury

All data expressed as median [interquartile range]. P values from Mann Whitney U test between groups.

Data available in a n=37,

b n=17,

c n=36,

d n=31,

e n=27,

f n=25,

g n=26.

h excluding subject M12 with undertreated primary hypogonadism

Abbreviations: bTBI: blast TBI, ND: not done, NHP: Nottingham Health Profile, nbTBI: non-blast TBI, SF-36: Short Form 36 Health Survey, TBI: traumatic brain

injury.

Note: For AGHDA, BDI-II, Epworth Sleepiness Scale, Pittsburgh Sleep Index and NHP higher score equals worse symptoms and quality of life; for SF-36 lower

score equals worse symptoms / quality of life.

Quality of Life / Symptom Assessment nbTBI All bTBIP value

nbTBI vs. bTBI

bTBI: No Pituitary

Dysfunction

bTBI: Pituitary

Dysfunction

P value

no pit dys vs. pit dys

n 38 18 h 12 h 6

Assessment of GH Deficiency in Adults (AGHDA) 9.5 [5.8-14.5] a 16.0 [4.0-18.5] b 0.22 14.0 [3.0-17.0] 17.5 [16.0-19.5] 0.10

Beck Depression Inventory Score (BDI-II) 11.0 [7.0-20.0] c 20.5 [4.0-24.5] 0.30 11.5 [1.8-21.8] 24.5 [20.3-26.3] 0.08

Epworth Sleepiness Scale 7.0 [2.0-12.0] d 7.0 [2.5-11.0] 0.89 6.0 [1.5-10.5] 10.0 [3.0-16.5] 0.25

Pittsburgh Sleep Index ND 10.0 [2.8-16.0] 4.5 [2.0-16.3] 12.0 [8.0-15.5] 0.37

NHP Energy Levels 39.0 [0-100] e 49.0 [18.0-100] 0.63 49.0 [0-94.0] 68.5 [24.0-100] 0.39

NHP Pain 0 [0-48] e 28.5 [9.7-52.5] 0.08 24.0 [14.0-51.0] 36.0 [0-54.5] 0.96

NHP Emotional Reactions 20.0 [10.0-47.0] e 32.5 [6.7-55.5] 0.71 15.0 [0-46.2] 54.0 [25.0-67.7] 0.10

NHP Sleep 22.0 [0-73.0] e 55.0 [0-100] 0.20 30.0 [0-93.2] 64.0 [31.7-100] 0.29

NHP Social Isolation 0 [0-45.0] e 21.0 [0-59.0] 0.52 0 [0-48.5] 29.0 [21.2-69.0] 0.13

NHP Physical Activity 0 [0-21.8] e 26.5 [11.0-42.0] 0.02 26.5 [13.2-42.0] 27.0 [0-60.7] 0.96

NHP Average 22.0 [5.4-41.0] e 41.5 [15.5-55.2] 0.09 28.5 [9.5-55.0] 48.5 [38.7-55.2] 0.25

NHP Daily Living Problems (0-7) 2.0 [0-5.0] f 5.0 [3.5-6.0] 0.04 4.5 [2.3-5.8] 4.5 [3.8-6.3] 0.62

SF-36 Physical functioning 85.0 [60.0-95.0] e 52.5 [27.5-81.3] 0.21 52.5 [41.3-58.8] 60.0 [27.5-88.8] 0.82

SF-36 Role limitations due to physical health 12.5 [0-62.5] g 12.5 [0-75.0] 0.94 25.0 [0-93.8] 0 [0-18.8] 0.10

SF-36 Role limitations due to emotional problems 67.0 [0-100] g 67.0 [24.8-100] 0.90 83.5 [33.0-100] 50.0 [0-100] 0.49

SF-36 Energy/Fatigue 50.0 [35-60] e 42.5 [33.8-66.3] 0.92 52.5 [36.3-70.0] 37.5 [26.3-47.0] 0.15

SF-36 Emotional well being 64.0 [52.0-80.0] e 60.0 [48.0-81.0] 0.57 68.0 [53.0-83.0] 58.0 [31.0-66.3] 0.34

SF-36 Social functioning 63.0 [38.0-75.0] e 50.0 [38.0-75.0] 0.66 56.5 [41.0-84.8] 44.0 [22.0-63.0] 0.18

SF-36 Pain 55.0 [33.0-88.0] e 45.0 [33.0-70.5] 0.68 68.0 [35.5-75.5] 33.0 [23.0-58.8] 0.21

SF-36 Health change 50.0 [38.0-75.0] e 32.5 [25.0-50.0] 0.06 25.0 [25.0-50.0] 45.0 [25.0-56.3] 0.49

SF-36 General health 50.0 [25.0-75.0] e 50.0 [25.0-61.3] 0.49 50.0 [27.5-63.8] 40.5 [18.8-61.3] 0.55

10

Table S6. Characteristics of soldiers with blast TBI

All data expressed as median [interquartile range] or n (%). P values from Mann Whitney U test or Fisher’s exact test between groups.

* for analgesia only, # on anti-epilepsy drug

Abbreviations: AIS: Abbreviated Injury Score, BMI: body mass index, GCS: Glasgow Coma Scale, GST: Glucagon stimulation test, ISS: Injury Severity Score, n/a:

not available, PTA: Post traumatic amnesia.

Subjects Age at TBI Age at GST Time since

TBIISS AIS Head AIS Chest AIS Abdo GCS PTA BMI at GST PTA >24 hrs

Limb

amputation Major organ damage

Skull/facial

fractureOpiate use

Antidepressant

useSeizures

Units / Maximum score Years Years Months 75 6 6 6 15 Days kg/m2

No pituitary dysfunction (n=13)

M02 36.3 37.6 15.2 20 4 0 2 3 1 25.4 No No No No Yes Yes No

M04 26.4 27.6 14.2 24 0 4 0 n/a 4 27.7 Yes No Lung/eye No No No No

M05 27.3 28.6 15.2 24 0 4 0 n/a 28 24.5 Yes No No No Yes No No

M09 19.0 19.6 6.7 45 4 0 2 n/a 4 n/a Yes Yes Perineum No Yes Yes * No

M11 19.3 20.9 16.6 25 5 0 0 3 84 26.6 Yes No No No No No Yes

M12 30.2 30.5 4.1 33 2 0 2 n/a 0 n/a No Yes Perineum No Yes Yes * No

M13 22.8 23.7 10.8 45 4 0 0 n/a 21 n/a Yes Yes Eye/Skin No No Yes * No

M15 26.4 26.8 4.1 45 4 0 2 15 0 n/a No Yes Eye/Skin/perineum No Yes Yes * No

M16 34.7 36.7 23.7 24 4 2 0 15 0 28.7 No No No No Yes Yes No

M17 26.6 28.0 16.6 9 3 0 0 14 0 23.6 No No No No No No No

M18 26.7 28.3 20.2 36 4 4 2 n/a 14 n/a Yes Yes Lung/colon/perineum No No No No

M19 26.6 27.7 13.6 9 3 0 0 n/a n/a n/a n/a Yes Skin No No No No

M20 30.9 32.2 15.4 9 3 0 0 n/a 2 29.4 Yes No No No No Yes No

median 26.6 28.0 15.2 24.0 4.0 0 0 3.0 26.6 7 (58.3%) 6 (46.1%) 7 (53.9%) 0 (0%) 6 (46.2%) 7 (53.8%) 1 (7.7%)

IQR [24.6-30.6] [25.3-31.4] [8.8-16.6] [14.5-40.5] [2.5-4.0] [0-3] [0-2] [0-19.3] [24.5-28.7]

Pituitary dysfunction (n=6)

M01 (PRL) 30.0 30.4 4.9 33 5 0 0 3 4 21.7 Yes No No Yes Yes Yes * Yes #

M03 (ACTH) 25.0 26.3 15.9 70 6 0 3 3 17 n/a Yes Yes Spleen/Liver Yes No Yes * No

M07 (GH) 34.3 36.2 21.9 38 5 3 0 3 7 26.7 Yes No Lung No No No No

M08 (ACTH/GH/Gn) 43.5 44.7 14.7 45 4 2 0 n/a 42 n/a Yes Yes No No Yes No No

M10 (PRL) 28.5 30.1 19.3 33 5 0 2 n/a 28 24.3 Yes No Eye/Liver/lung Yes No No No

M14 (GH) 26.1 27.7 19.6 9 0 1 0 3 14 33.7 Yes No Skin No Yes Yes No

median 29.3 30.3 17.6 35.5 5 0.5 0 15.5 25.5 6 (100%) 2 (33.3%) 4 (66.7%) 3 (50.0%) 3 (50%) 3 (50%) 1 (16.7%)

IQR [25.8-36.6] [27.4-38.3] [12.3-20.2] [27.0-51.3] [3.0-5.3] [0-2.3] [0-2.3] [6.3-31.5] [22.4-32.0]

P value 0.48 0.32 0.32 0.24 0.06 0.83 0.97 0.10 0.79 0.11 1.00 1.00 0.02 1.00 1.00 1.00

11

Table S7. Medications used by soldiers with blast TBI

Abbreviations: MST: morphine sulphate

Subjects Medications

No pituitary dysfunction (n=13)

M02 Diclofenac, Sertraline, Tramadol

M04 Co-codamol

M05 Diclofenac, Tramadol

M09 Amitriptyline, MST, Nebido, Pregabalin

M11 None

M12 Amitriptyline, Diclofenac, Nebido, Pregabalin, Ranitidine, Sildenafil, Tramadol

M13 Amitriptyline, Baclofen, Pregabalin

M15 Amitriptyline, Nebido, Pregabalin, Tramadol

M16 Mirtazepine, Paracetamol, Pregabalin, Tramadol, Zopiclone

M17 None

M18 Nebido

M19 Diclofenac, Pregabalin, Ranitidine

M20 Sertraline, Zopiclone

Pituitary dysfunction (n=6)

M01 (PRL) Amitriptyline, Diclofenac, MST, Phenytoin

M03 (ACTH) Amitriptyline, Erythromycin, Gabapentin

M07 (GH) None

M08 (ACTH/GH/Gn) Diclofenac, Lansoprazole, MST, Paracetamol, Pregabalin, Tramadol

M10 (PRL) Betnovate ointment, Co-codamol

M14 (GH) Amitriptyline, Diclofenac, Fluoxetine, Mirtazepine, MST, Paracetamol, Pregabalin, Salbutamol inhaler, Temazepam, Zopiclone

12

SUPPLEMENTAL RESULTS

Non-pituitary endocrine diagnoses in bTBI and nbTBI cohorts

Other non-pituitary endocrine disorders were diagnosed in both groups. Primary hypogonadism

due to perineum/testicular blast injury had been found in 4 out of 19 soldiers (21.2%), none of

whom had pituitary dysfunction (Table 2 and S1). Although at the time of our assessment all these

subjects were already on testosterone replacement, 3 had documented increased gonadotophins

before its initiation (Table S1). One of these (M12) was under-replaced with testosterone at the

time of assessment. A high prevalence of perineal blast injury has previously been reported in

soldiers exposed to IED (Mossadegh et al., 2012). One control patient with nbTBI had a pre-

existing diagnosis of primary hypothyroidism, and another had previously undiagnosed primary

hypogonadism of unknown cause unrelated to their nbTBI.

IGF-I levels in bTBI patients with GH deficiency

IGF-I levels were within the normal range in all those soldiers with GH deficiency. When comparing

those soldiers with bTBI who had GH deficiency (n=3) to those without GH deficiency (n=16),

absolute IGF-I levels tended to be lower in those with than without GH deficiency (median [IQR]

18.2 [16.7-22.3] vs. 27.1 (19.9-31.6], P=0.11). However IGF-I relative to median of age-related

reference range were similar between groups (0.66 [0.60-0.73] vs. 0.79 [0.63-1.00], P=0.40) (Table

S1).

Symptoms, quality of life and cognitive function

In our cohort of soldiers with bTBI, subjective symptoms included worsening of their memory

(70%), changes in mood (70%), difficulty concentrating (65%), difficulty sleeping (55%), headaches

(45%), and dizziness (30%).

Consistent with their higher prevalence of polytrauma and amputations, the soldiers with bTBI had

significantly worse scores for physical activity (P=0.02) and daily living problems (P=0.04) from the

Nottingham Health Profile (NHP) questionnaire, with a tendency for worse NHP pain scores

(P=0.08) and change in health from the Short Form-36 (SF-36) quality of life questionnaire

13

(P=0.06), than the control nbTBI group (Table S5). However there were no significant differences

in measures of depression and emotional well-being (from Beck Depression Inventory-II), NHP and

SF-36 questionnaires) between the bTBI and nbTBI groups (P=0.30-0.71) (Table S5).

In the bTBI group, soldiers with pituitary dysfunction had trends towards worse measures of QoL

and symptom scores in several domains compared to those without pituitary dysfunction (Table

S5). Soldiers after bTBI with pituitary dysfunction had trends for higher AGHDA QoL score

(P=0.10), worse scores for emotional reactions (NHP, P=0.10), social isolation (NHP, P=0.13), role

limitations due to physical health (SF-36, P=0.10), energy/fatigue (SF-36, P=0.15), and social

functioning (SF-36, P=0.18), and higher depression scores (BDI-II, P=0.10), though none had

symptoms suggesting severe depression (all scores <28/63).

Interpretation of metyrapone test

Although the metyrapone test is not a commonly used test for ACTH deficiency (Grossman 2010),

it was only needed for the confirmatory diagnosis in one soldier (M03). Furthermore that subject

also had very low cortisol levels throughout their day curve ≤50 nmol/L (≤1.81 μg/dL) confirming

the diagnosis of ACTH deficiency. The second soldier with ACTH deficiency (M10) failed their

cortisol response to insulin-induced hypoglycemia (peak 268 nmol/L), and also had low cortisol

levels (<100 nmol/L, <3.62 μg/dL) at 1200h on their day curve supporting the diagnosis. Other

soldiers who initially had low cortisol responses to glucagon stimulation, subsequently had ACTH

deficiency excluded on the basis of normal responses to ITT (M02) or Metyrapone test (M10), but

both also had subsequent high basal morning cortisol levels (M02, M10, >400 nmol/L, 14.50

μg/dL).

Previous studies comparing the metyrapone test to more commonly used tests for ACTH

deficiency have demonstrated the metyrapone test to have specificity, sensitivity and concordance

(accuracy) rates of 77-100%, 64-89%, 74-84% (n=17-32) and 86, 91, 87% (n=87) with the ITT and

ACTH stimulation test respectively (Fiad et al. 1994; Courtney et al. 2000; Giordano et al. 2008).

Furthermore in a recent audit of patients from our endocrine clinics suspected of having ACTH

14

deficiency (n=24, excluding soldiers with bTBI from this study), we have found an overall 92%

concordance rate between results of a metyrapone test, and the ACTH stimulation test (n=12,

normal response >480 nmol/L or 17.40 μg/dL, using alignment of the previous 550 nmol/L cut-off to

the new Architect i2000 assay) or ITT (n=13) (unpublished observations). In this analysis, all

patients failing the metyrapone test (n=5) also failed an ITT. The overall specificity for the

metyrapone test in diagnosing ACTH deficiency was 100% and sensitivity was 71% (unpublished

observations).

15

SUPPLEMENTAL METHODS

Recruitment

Ethical approval was granted by the Ealing and West London Hospitals Research Ethics

Committee. Studies were performed according to the Declaration of Helsinki and all soldiers gave

informed written consent.

Inclusion of a military combat nbTBI group would have been a useful in addition to the civilian

nbTBI group to control for active military service in an identical theatre. However in UK soldiers

experiencing nbTBI in Afghanistan, the majority are due to gunshot wounds that are either fatal or

complicated by penetrating brain injury often requiring surgery. The lower prevalence of military

non-penetrating nbTBI, primarily due to road traffic accidents, precluded endocrine assessment of

a sufficient number of such soldiers to be included in this study.

Both bTBI and nbTBI subjects had clinical assessment, calculation of their Abbreviated Injury

Scores (AIS) for each body region including brain, and total Injury Severity Score (ISS) (Baker et

al. 1974; Hawley 1996), and completed quality of life (QoL) and symptom questionnaires:

Assessment of Growth Hormone Deficiency in Adults (QoL-AGHDA); Beck Depression Inventory-II

(BDI-II); Nottingham Health Profile (NHP); Short Form 36 Health Survey (SF-36), Pittsburgh Sleep

Quality Index and Epworth Sleepiness Scale (Hunt et al. 1985; Buysse et al. 1989; Johns 1991;

Ware & Sherbourne 1992; Beck et al. 1996; McKenna et al. 1999). Soldiers were excluded if they

had needed massive blood transfusion so as to exclude pituitary dysfunction secondary to

hypovolemic shock (Stainsby et al. 2006).

Endocrine Testing

Endocrine assessment included baseline measurement of serum anterior pituitary hormones: TSH,

free T4, free T3, prolactin, FSH, LH, testosterone (Abbott Architect Ci8200), ACTH, cortisol, GH,

IGF-I (Immulite® 2000) and sex hormone binding globulin (SHBG). Free androgen index was

calculated as 100 x total testosterone / SHBG.

16

A diagnosis of hyperprolactinemia was made on the basis of two consecutively raised prolactin

readings (above upper reference range, Table 1) and a negative macroprolactin, an immunological

artefact leading to misdiagnosis of hyperprolactinemia (assessed by PEG precipitation) (Smith et

al. 2007). Subjects who met these criteria had MRI of the pituitary including gadolinium contrast to

rule out an incidental pituitary tumour.

A diagnosis of gonadotrophin deficiency was made on the basis of a low morning testosterone <10

nmol/L (<2.9 ng/mL) with low or non-elevated LH (NR 1.7-12.0 IU/L) and FSH (NR 1.7-8.0 IU/L). If

sex hormone binding globulin (SHBG) was low (<15 nmol/L), then FAI needed to be <30 for the

diagnosis. Primary hypogonadism was defined as a low morning testosterone or FAI with elevated

FSH and/or LH.

Growth hormone (GH) deficiency was defined as failure on 2 dynamic endocrine tests performed in

the morning: (i) Glucagon Stimulation Test (GST) used as initial screening test and (ii) a

confirmatory 2nd line test, either the GHRH-Arginine Test or an Insulin Tolerance Test (ITT).

Similarly, a diagnosis of ACTH deficiency was made on the basis of failure on 2 dynamic endocrine

tests performed in the morning: (i) a GST, and (ii) an ITT or an overnight Metyrapone Stimulation

Test (MST). A 5 point Cortisol Day Curve (CDC) was also used to help confirm or exclude ACTH

deficiency, and assess the need for maintenance hydrocortisone replacement as opposed to just

during intercurrent illness.

An ITT was not routinely performed because of the prevalence of relative and absolute

contraindications in this population. In our cohort 10.5% of soldiers after bTBI and 10.3% of

controls after nbTBI had an absolute contraindication (history of seizures, ischemic heart disease,

cardiac arrhythmias, abnormal ECG), whilst an additional 21.1% and 53.8% had a relative

contraindication (intra-cerebral contusion, intra-cranial hemorrhage). If further confirmatory testing

was required because of equivocal findings on the second dynamic test (e.g. difficulty calculating

17

BMI in soldiers with amputations), and no contraindications were present, an ITT was carried out in

addition to the glucagon test and GHRH-Arginine or metyrapone test.

Diabetes insipidus was screened for on the basis of symptoms (polyuria and polydipsia) and

measurement of paired random clinic urine and plasma osmolalities. If clinically indicated, a Water

Deprivation Test was performed (n=6 controls with nbTBI, n=1 soldier with bTBI).

All dynamic endocrine tests were carried out in an in-patient facility at Charing Cross Hospital,

London or St. Mary’s Hospital, London. A summary of the algorithm used to define pituitary

dysfunction is shown in Table 1.

Glucagon Stimulation Test (GST)

Following an overnight fast, patients had basal blood samples. Glucagon (GlucaGen™, Novo

Nordisk Pharmaceuticals, Crawley, UK 1 mg, or 1.5 mg if weight >90 kg) was administered

intramuscularly. Blood samples for glucose, serum cortisol and GH were taken at 90, 120, 150 and

180 minutes after glucagon administration from an intravenous (IV) cannula. The majority of

subjects (89% soldiers and 70% controls) also had samples taken at 210 and 240 minutes. An

abnormal response was defined as a peak GH <5 μg/L and cortisol <350 nmol/L (<12.7 μg/dL)

during the test (Yuen et al. 2009; Cegla et al. 2012). Subjects who failed to reach these thresholds

underwent at least one additional confirmatory dynamic test.

The method for cortisol determination was changed in August 2010 from the Immulite® 2000

assay (Siemens) to a chemiluminescence immunoassay with the Architect i2000 (Abbott, UK). To

assure comparability, quality controls and linear regression analysis were performed (data not

shown) and results from the Immulite assay were aligned with the Architect i2000 assay. The

Architect assay has coefficients of variation <10% for cortisol levels of 83–967 nmol/L (3.0-35.0

μg/dL).

18

GHRH-Arginine Test

Following an overnight fast, patients had blood samples taken for GH and IGF-I measurement at 0

minutes. GHRH (Somatorelin, Ferring) 1μg/kg was given as a bolus IV injection into one arm

followed by the IV infusion of 0.5g/kg L-arginine monohydrochloride (Stockport Pharmaceuticals)

as a 10% solution (30g/300mL up to a maximum of 30g) in normal saline over 30 minutes (Colao

et al., 2009). Further blood samples for GH estimation were taken at +30, 60, 90, 120 and 150

minutes after the start of the arginine infusion.

GH cut offs to confirm GH deficiency varied according to age and BMI. For age groups 15-25 years

old, 26-65 years old and older than 65 years, GH cut-offs were respectively <15.6, <11.7, and <8.5

μg/L, <11.8, <8.1, and <5.5 μg/L, and <9.2, <6.1, and <4.0 μg/L, respectively, in lean

(BMI<25.0kg/m2), overweight (BMI 25.0-30.0 kg/m2) and obese (BMI>30.0 kg/m2) subjects (Colao

et al. 2009). If amputations precluded accurate determination of BMI then cut-offs in the overweight

range were used.

Insulin Tolerance Test (ITT)

Following an overnight fast, basal blood samples were taken and IV insulin Actrapid (NovoNordisk)

administered (0.15 U/kg). Blood samples were taken for GH, cortisol and glucose at 0, 30, 60, 90,

and 120 mins. Blood glucose was also measured simultaneously. Once adequate hypoglycemia

(<2.2 mmol/L, <39.6 mg/dL) was achieved, hypoglycemia was reversed with oral glucose and at

least two further blood specimens were taken before test completion.

Abnormal cortisol response was defined as peak cortisol of <450 nmol/L (<16.3 μg/dL) providing

adequate hypoglycemia was achieved (using alignment of the previous 500 nmol/L cut-off to the

new Architect i2000 assay). Severe GH deficiency was defined as a peak GH <3 μg/L (Plumpton &

Besser 1969; Fish et al. 1986; Molitch et al. 2011).

Cortisol Day Curve

Blood samples were taken from an IV cannula for serum cortisol estimation at 0900h, 1200h,

19

1500h, 1800h and 2100h (Immulite ® 2000 assay (Siemens) or Architect i2000 (Abbott, UK), and

plasma ACTH at 0900h. Results helped confirm (cortisol <100 nmol/L or 3.62 μg/dL at 0900 or

1200h), or exclude (cortisol >400 nmol/L or 14.50 μg/dL at 0900h) ACTH deficiency, and assess

the need for maintenance hydrocortisone replacement as opposed to just during intercurrent illness

(Grossman 2010).

Metyrapone Stimulation Test

Patients were given oral metyrapone (Metopirone™, Alliance Pharmaceuticals, Chippenham, UK)

(30 mg/kg), at midnight with a snack, according to their body weight (<70 kg 2.0g, 70-90 kg 2.5g,

>90kg 3.0g) (Steiner et al. 1994; Cegla et al. 2012). At 0900h the following morning, blood samples

were taken for serum cortisol, 11-DOC (Biosource, Oxford Biosystems, UK) and plasma ACTH

(Immulite® 2000, Siemens). Hydrocortisone 10 mg was given orally to counteract hypocortisolism

and the patients were discharged.

Metyrapone causes inhibition of 11 β-hydroxylase (used in the conversion of 11-deoxycortisol to

cortisol) and cortisol suppression to <200 nmol/L (7.25 μg/dL) is the desired threshold to stimulate

ACTH drive. Subjects were considered to be ACTH sufficient if 11-DOC was >200 nmol/L or, if the

11-DOC was unavailable, if ACTH >60 ng/L (Steiner et al. 1994; Cegla et al. 2012).

Water Deprivation Test

This was carried out in two stages on non-fasted subjects (Vokes & Robertson 1988).

In Stage 1, patients drank no fluid from 0830-1630h. Weight and urine volume (after urine passed

and discarded at t=0) were recorded hourly. The test was stopped if >3% weight was lost. Urine

specimens were taken for osmolality from the total hourly sample passes over 0830-0930h (U1),

1130-1230h (U2), 1430-1530h (U3), 1530-1630h (U4). Blood samples were taken for osmolality

and plasma sodium at 0900h (P1), 1200h (P2), 1500h (P3) and 1600h (P4).

In Stage 2, at 1630 hrs following the dehydration stage, Desmopressin (DDAVP 2μg IM or 20μg

intra-nasally) was administered. Urine volumes were recorded and urine specimens for osmolality

20

measurement were taken every hour until test completion at 2030h.

Central diabetes insipidus was defined as plasma concentration to >300 mosmol/kg with

inappropriately hypotonic urine (U3:P3 or U4:P4 ≤1.9) or urine osmolality <350 mosmol/kg. In

addition, urine was required to concentrate to >150% of previous highest value following DDAVP

administration.

Neuropsychological Assessments

Each soldier completed a standardized neuropsychological test battery previously shown to be

sensitive to cognitive impairment associated with traumatic brain injury (Kinnunen et al. 2011). The

cognitive functions of specific interest were indexed by: (i) current verbal and non-verbal reasoning

ability via the Wechsler Abbreviated Scale of Intelligence Similarities and Matrix Reasoning

subtests (Wechsler 1999); (ii) associative learning and memory via the immediate recall score on

the People Test from the Doors and People Test (Baddeley 2011); (iii) the executive functions of

set-shifting, inhibitory control, cognitive flexibility and word generation fluency via the Trail Making

Test alternating-switch cost index (time to complete alternating letter and number Trails B - time to

complete numbers only Trail A) and two indices from the Delis-Kaplan Executive Function System

(Reitan 1958; Delis et al. 2001), namely the inhibition/switching minus baseline score from the

Colour-Word subtest (high scores indicating poor performance) and the total score on Letter

Fluency; and (iv) information processing speed via the median reaction time for accurate

responses on a simple computerized choice reaction task (Kinnunen et al. 2011). The Wechsler

Test of Adult Reading (WTAR) was also administered as a measure of pre-morbid intelligence

(Green et al. 2008).

Structural Imaging

Each soldier had standard high-resolution T1 and gradient-echo (T2*) (1.75x1.75x2mm3) imaging

to assess focal brain injury and evidence of microbleeds, superficial siderosis, presence and

location of contusions and gross pituitary injury. All structural MR scans were reviewed by a single

experienced consultant neuroradiologist. Contusion volume was calculated by converting the T1

21

images into standard 1mm MNI brain space using FLIRT (FMRIB, University of Oxford, UK) and

manually drawing a mask in the z plane.

MRI was performed on 3T Achieva scanner (Philips Medical Systems, Netherlands) using an 8

channel head coil. The T1 and T2*-weighted images were obtained prior to DTI. For DTI, diffusion-

weighted volumes with gradients applied in 16 non-collinear directions were collected in each of

the four DTI runs, resulting in a total of 64 directions. The following parameters were used: 73

contiguous slices, slice thickness 2mm, field of view 224mm, matrix 128 x 128 (voxel size

1.75x1.75x2 mm3), b value 1000 and four images with no diffusion weighting (b=0s/mm2).

The images were registered to the b0 image by affine transformations to minimize distortion due to

motion and eddy currents and then brain-extracted using Brain Extraction Tool (Smith 2002) from

the FMRIB Software Library image processing toolbox (Smith et al. 2004; Woolrich et al. 2009).

Fractional anisotropy (FA) maps were generated using the Diffusion Toolbox (Behrens et al. 2003).

DTI Analysis

DTI analysis used TBSS and non-parametric permutation based statistics for whole brain and

region of interest (ROI) analysis (FMRIB software, FSL, University of Oxford, UK).

Voxelwise analysis of the fractional anisotropy, was carried out using TBSS in the FMRIB Software

Library (Smith et al. 2004; Smith et al. 2006). Image analysis using TBSS involved a number of

steps: (i) non-linear alignment of all subjects’ FA images into common FMRIB58 FA template

space; (ii) affine-transformation of the aligned images into standard MNI152 1mm space; (iii)

averaging of the aligned FA images to create a 4D mean FA image; (iv) thinning of the mean FA

image to create a mean FA ‘skeleton’ representing the centre of all white matter tracts, and in this

way removing partial-volume confounds; and (v) thresholding of the FA skeleton at FA 0.2 to

suppress areas of extremely low mean FA and exclude those with considerable inter-individual

variability. Non-parametric permutation-based statistics were employed using randomize with

threshold-free cluster enhancement and 5000 permutations (Nichols & Holmes 2002; Smith &

22

Nichols 2009). A threshold of P<0.05 was then applied on the results, corrected for multiple

comparisons. Age was included as a covariate of no interest in all TBSS analyses.

Regions of interest (ROI) were defined using the John Hopkins University (JHU) white matter atlas.

We chose 10 areas that represented white matter regions throughout the whole brain and have

been shown to be damaged in nbTBI as well as mild bTBI (Kinnunen et al. 2011; MacDonald et al.

2011). These regions were: anterior and posterior internal capsules, cingulum, body/genu and

splenium of the corpus callosum, cerebral peduncles, middle cerebellar peduncles, and uncinate

fasciuli (Fig. S1). In addition a cerebellum ROI mask was drawn manually and an orbitofrontal

white matter ROI mask made using the Washington University, St Louis criteria from the standard

MNI152 1mm T1 brain (MacDonald et al. 2011). A repeated measures ANOVA was performed to

assess the overall significance effect of pituitary dysfunction on FA, including group, ROI and

group x ROI interaction as independent variables, with post-hoc 2-tailed t-tests for comparison of

FA in individual ROIs between groups.

23

SUPPLEMENTAL REFERENCES

Baddeley A. (2011) Doors and people test: a test of visual and verbal recall and recognition.

Thames Valley Test Company, Bury St Edmunds, Suffolk, UK.

Baker SP, O'Neill B, Haddon W, Jr., Long WB. (1974) The injury severity score: a method for

describing patients with multiple injuries and evaluating emergency care. J Trauma 14: 187-96.

Beck AT, Steer RA, Ball R, Ranieri W. (1996) Comparison of Beck Depression Inventories -IA and

-II in psychiatric outpatients. J Pers Assess 67: 588-97.

Behrens TE, Woolrich MW, Jenkinson M, Johansen-Berg H, Nunes RG, Clare S, Matthews PM,

Brady JM, Smith SM. (2003) Characterization and propagation of uncertainty in diffusion-weighted

MR imaging. Magn Reson Med 50: 1077-88.

Buysse DJ, Reynolds CF, Monk TH, Berman SR, Kupfer DJ. (1989) The Pittsburgh Sleep Quality

Index: a new instrument for psychiatric practice and research. Psychiatry Res 28: 193-213.

Cegla J, Jones B, Seyani L, Papadoulou D, Wynne K, Martin NM, Meeran K, Chapman R,

Donaldson M, Goldstone AP, Tan T. (2012) Comparison of the overnight metyrapone and

glucagon stimulation tests in the assessment of secondary hypoadrenalism. Clin Endocrinol (Oxf)

doi: 10.1111/cen.12043.

Colao A, Di SC, Savastano S, Rota F, Savanelli MC, Aimaretti G, Lombardi G. (2009) A

reappraisal of diagnosing GH deficiency in adults: role of gender, age, waist circumference, and

body mass index. J Clin Endocrinol Metab 94: 4414-22.

Courtney CH, McAllister AS, McCance DR, Hadden DR, Leslie H, Sheridan B, Atkinson AB. (2000)

The insulin hypoglycaemia and overnight metyrapone tests in the assessment of the hypothalamic-

pituitary-adrenal axis following pituitary surgery. Clin Endocrinol (Oxf) 53: 309-12.

Delis DC, Kaplan E, Kramer JH. (2001) Delis-Kaplan Executive Function System. The

Psychological Corporation, San Antonio, Texas, USA.

24

Fiad TM, Kirby JM, Cunningham SK, McKenna TJ. (1994) The overnight single-dose metyrapone

test is a simple and reliable index of the hypothalamic-pituitary-adrenal axis. Clin Endocrinol (Oxf)

40: 603-9.

Fish HR, Chernow B, O'Brian JT. (1986) Endocrine and neurophysiologic responses of the pituitary

to insulin-induced hypoglycemia: a review. Metabolism 35: 763-80.

Giordano R, Picu A, Bonelli L, Balbo M, Berardelli R, Marinazzo E, Corneli G, Ghigo E, Arvat E.

(2008) Hypothalamus-pituitary-adrenal axis evaluation in patients with hypothalamo-pituitary

disorders: comparison of different provocative tests. Clin Endocrinol (Oxf) 68: 935-41.

Green RE, Melo B, Christensen B, Ngo LA, Monette G, Bradbury C. (2008) Measuring premorbid

IQ in traumatic brain injury: an examination of the validity of the Wechsler Test of Adult Reading

(WTAR). J Clin Exp Neuropsychol 30: 163-72.

Grossman AB. (2010) Clinical Review: The diagnosis and management of central hypoadrenalism.

J Clin Endocrinol Metab 95: 4855-63.

Hawley A. (1996) Trauma management on the battlefield: a modern approach. J R Army Med

Corps 142: 120-5.

Hunt SM, McEwen J, McKenna SP. (1985) Measuring health status: a new tool for clinicians and

epidemiologists. J R Coll Gen Pract 35: 185-8.

Johns MW. (1991) A new method for measuring daytime sleepiness: the Epworth sleepiness scale.

Sleep 14: 540-5.

Kinnunen KM, Greenwood R, Powell JH, Leech R, Hawkins PC, Bonnelle V, Patel MC, Counsell

SJ, Sharp DJ. (2011) White matter damage and cognitive impairment after traumatic brain injury.

Brain 134: 449-63.

MacDonald CL, Johnson AM, Cooper D, Nelson EC, Werner NJ, Shimony JS, Snyder AZ, Raichle

ME, Witherow JR, Fang R, Flaherty SF, Brody DL. (2011) Detection of blast-related traumatic brain

injury in U.S. military personnel. N Engl J Med 364: 2091-100.

25

McKenna SP, Doward LC, Alonso J, Kohlmann T, Niero M, Prieto L, Wiren L. (1999) The QoL-

AGHDA: an instrument for the assessment of quality of life in adults with growth hormone

deficiency. Qual Life Res 8: 373-83.

Molitch ME, Clemmons DR, Malozowski S, Merriam GR, Vance ML. (2011) Evaluation and

treatment of adult growth hormone deficiency: an Endocrine Society clinical practice guideline. J

Clin Endocrinol Metab 96: 1587-609.

Nichols TE, Holmes AP. (2002) Nonparametric permutation tests for functional neuroimaging: a

primer with examples. Hum Brain Mapp 15: 1-25.

Plumpton FS, Besser GM. (1969) The adrenocortical response to surgery and insulin-induced

hypoglycaemia in corticosteroid-treated and normal subjects. Br J Surg 56: 216-9.

Reitan R. (1958) The validity of the Trail Making Test as an indicator of organic brain damage.

Percept Mot Skills 8: 276.

Smith SM. (2002) Fast robust automated brain extraction. Hum Brain Mapp 17: 143-55.

Smith SM, Jenkinson M, Johansen-Berg H, Rueckert D, Nichols TE, Mackay CE, Watkins KE,

Ciccarelli O, Cader MZ, Matthews PM, Behrens TE. (2006) Tract-based spatial statistics: voxelwise

analysis of multi-subject diffusion data. Neuroimage 31: 1487-505.

Smith SM, Jenkinson M, Woolrich MW, Beckmann CF, Behrens TE, Johansen-Berg H, Bannister

PR, De LM, Drobnjak I, Flitney DE, Niazy RK, Saunders J, Vickers J, Zhang Y, De SN, Brady JM,

Matthews PM. (2004) Advances in functional and structural MR image analysis and

implementation as FSL. Neuroimage 23 Suppl 1: S208-S219.

Smith SM, Nichols TE. (2009) Threshold-free cluster enhancement: addressing problems of

smoothing, threshold dependence and localisation in cluster inference. Neuroimage 44: 83-98.

Smith TP, Kavanagh L, Healy ML, McKenna TJ. (2007) Technology insight: measuring prolactin in

clinical samples. Nat Clin Pract Endocrinol Metab 3: 279-89.

26

Stainsby D, MacLennan S, Thomas D, Isaac J, Hamilton PJ. (2006) Guidelines on the

management of massive blood loss. Br J Haematol 135: 634-41.

Steiner H, Bahr V, Exner P, Oelkers PW. (1994) Pituitary function tests: comparison of ACTH and

11-deoxy-cortisol responses in the metyrapone test and with the insulin hypoglycemia test. Exp

Clin Endocrinol 102: 33-8.

Vokes TJ, Robertson GL. (1988) Disorders of antidiuretic hormone. Endocrinol Metab Clin North

Am 17: 281-99.

Ware JE, Sherbourne CD. (1992) The MOS 36-item short-form health survey (SF-36). I.

Conceptual framework and item selection. Med Care 30: 473-83.

Wechsler D. (1999) WASI: Wechsler Abbreviated Scale of Intelligence. The Psychological

Corporation, San Antonio, Texas, USA.

Woolrich MW, Jbabdi S, Patenaude B, Chappell M, Makni S, Behrens T, Beckmann C, Jenkinson

M, Smith SM. (2009) Bayesian analysis of neuroimaging data in FSL. Neuroimage 45 Suppl 1:

S173-S186.

Yuen KC, Biller BM, Molitch ME, Cook DM. (2009) Clinical review: Is lack of recombinant growth

hormone (GH)-releasing hormone in the United States a setback or time to consider glucagon

testing for adult GH deficiency? J Clin Endocrinol Metab 94: 2702-7.

Pituitary Dysfunction after Blast Traumatic Brain Injury: The UK BIOSAP Study

Documents

Transcript of Pituitary Dysfunction after Blast Traumatic Brain Injury: The UK BIOSAP Study