Persistent hypogonadism influences estradiol synthesis, cognition and outcome in males after severe TBI

Abstract

Objective: Acute hypogonadotropic hypogonadism (AHH) occurs frequently after TBI, as does chronic hypogonadotropic hypogonadism. However, AHH and persistent hypogonadotropic hypogonadism (PHH) after TBI are not well studied. The objective of this study was to characterize longitudinal hormone profiles and the impact of AHH and PHH on outcome.

Methods: In this prospective cohort study, men with severe TBI (n = 38) had serum gonadal and gonadotropic hormones measured during weeks 1–52 post-injury. AHH, PHH and/or early resolving hypogonadotropic hypogonadism (ERHH) were based on temporal hormone assessments. PHH and hormone profiles were then compared to multiple outcome measures 6–12 months post-TBI.

Results: AHH affected 100% of the population, while 37% subsequently developed PHH. Acute testosterone (TEST) and estradiol/testosterone (E2/TEST) ratios were associated with PHH and outcome. Over time, post-acute TEST and E2 levels for the ERHH group approached normal range, while levels for the PHH group remained low. Post-acute gonadotrophin levels were within the normal range for both groups. PHH, along with lower post-acute TEST and E2 profiles, was associated with worse functional and cognitive outcomes at 6 and 12 months post-injury.

Conclusions: These results support screening for post-acute secondary hypogonadism and further research to assess the mechanisms underlying PHH and associated functional and cognitive deficits.

Introduction

Traumatic brain injury (TBI) is one of the leading causes of death and life-long disability for people in the US [1]. There is a significantly higher incidence of TBI in men than women, with 78.8% of TBI injuries involving men [2]. The estimated number of Americans living with TBI is 1.7 million [3], many of whom have persistent symptoms and functional deficits.

Pituitary-related complications frequently occur after TBI. This may in part be attributed to the susceptibility of the hypothalamus and anterior pituitary to damage during head trauma and other factors that subsequently suppress the hypothalamic-pituitary-gonadal (HPG) axis. Damage to the hypothalamic-pituitary system can be caused by vascular injury, involving traumatic rupture of the hypophysial portal veins of the pituitary stalk, which nourishes 90% of the anterior pituitary [4]. This mechanism is supported by post-mortem studies showing ischemic necrosis and haemorrhage in the pituitary gland and its stalk in individuals with very severe injuries who died soon after TBI. Other possible mechanisms of hypothalamic-pituitary injury are compression secondary to oedema, skull fracture, increased intracranial pressure, hypoxic insult or direct mechanical injury [5].

In addition to the findings above, changes to hypothalamic-pituitary functioning may occur in some individuals with TBI, even without direct anatomical evidence of injury involving hypothalamus and pituitary structures. For instance, hypopituitarism can occur in the setting of mild TBI, for which there is no clinical radiographic evidence of any injury. Additionally, the stress response, which occurs shortly after acute illness or critical injury, leads to nearly uniform functional deficits in the HPG axis after severe TBI, independent of direct injury [6–10]. This critical illness stress response can also lead to peripheral aromatase activation and elevated estradiol (E2) levels, which are associated with poor long-term prognosis and mortality [11–14].

Hypogonadotropic (secondary) hypogonadism is the most common presentation for both early and chronic post-traumatic hypopituitarism and it is generally defined in men as absent or decreased testicular production of testosterone (TEST) that is attributable to inadequate pituitary hormone production, particularly leutenizing hormone (LH). The prevalence of chronic hypopituitarism varies considerably from study to study, ranging from 1.4–29% [5], [15], [16]. This variability in the reported incidence of hypogonadism can be partly attributed to study differences in how trauma severity and timing of hormonal assessment post-injury are evaluated, but the variation may also be dependent on criteria used for diagnosis. For example, some studies have assessed anterior pituitary hormone and gonadotropin deficiencies for only one or two fixed time points post-injury, which may not fully capture hormonal changes; in other reports, hormone function has been assessed in the context of hypogonadism categorization at variable time points post-injury [5], [17], without accounting for the evolution of hormone levels over the acute and chronic time course post-injury or the persistence of hormone deficiencies over time.

In general, hypogonadism in men is associated with decreased life quality, fatigue, mood impairment, insomnia, osteoporosis and impaired sexual function. However, it is not uncommon for hypogonadism to be a sub-clinical condition, with few or none of the overt clinical manifestations, in which case it may only be identified only by hormonal tests [18]. Of significant interest is the fact that, in studies involving the elderly and in those with dementia, decreased cognitive functioning has been linked with decreased TEST levels [19], [20]. Based on these findings, screening for post-traumatic hypogonadism may be important in men, even in the absence of clinical symptoms.

Of the published work in post-traumatic hypogonadism, most studies have involved only a limited examination of hypogonadism relationships on injury severity and outcome [5], [21]. When examining literature assessing TBI-mediated hypogonadism and outcome, one study showed large positive gains in the Functional Independence Measure (FIM) as associated with increased TEST measured at both admission and discharge in a population receiving inpatient rehabilitation [22]. No studies have assessed both TEST and pituitary hormones profiles over time in relation to long-term TBI outcome. Importantly, no studies have extensively assessed how chronic E2 levels over time are associated with long-term outcome in men, even though E2 is the byproduct of TEST metabolism. However, TEST replacement in men can improve cognition in older men with hypogonadism, in part through aromatization to E2 [23].

In order to understand the evolution and nature of TBI-mediated hypogonadism, its effects on cognition and outcome and the mechanisms by which these effects occur, this exploratory study examined a group of male survivors with severe TBI and assessed both acute and chronic serum E2, TEST and pituitary hormone production over the first 12 months post-injury. Relationships between the clinically relevant designation of hypogonadism, gonadal hormone and gonadotropin levels and a battery of functional outcome and cognition measures at 6 and 12 months post-injury were also assessed. It was hypothesized that persistent hypogonadism and the associated low hormone levels over time would be correlated with worse outcomes. The data demonstrated uniform development of acute hypogonadotropic hypogonadism (AHH) and indicated that, compared to those with early resolving hypogonadotropic hypogonadism (ERHH), persistent hypogonadotropic hypogonadism (PHH) occurs frequently after TBI and is associated with a low TEST and E2 state as well as worse outcomes. Given the established role of E2 in mediating neuroplasticity through various mechanisms, the relative decreases in E2 observed in men with PHH may be one factor that contributes to worse functional and cognitive outcomes observed in this group [24–27].

Methods

Study design and description of subjects

This study was approved by the Institutional Review Board at The University of Pittsburgh.

Thirty-eight male survivors with severe TBI were enrolled in this prospective longitudinal cohort study. To be included for analysis, subjects were within the age range of 16–70 years old, had a severe TBI based on a GCS ≤ 8 with positive findings on head CT and required an extraventricular drainage catheter (EVD) for intracranial pressure monitoring and management. People with a history of pituitary or hypothalamic tumour, a condition requiring orchiectomy and/or LH suppression drugs or untreated thyroid disease were excluded from the study. Furthermore, subjects included in the study cohort had to have completed all evaluated outcomes or, if the subjects were cognitively unable to complete the tasks, they were assigned the worst possible scores as applicable. Because some prospective studies investigating early cognitive recovery following moderate and severe injury have found that most recovery generally occurs within the first 6–18 months of injury [28–30] and more gradual improvement may continue subsequently [31–33], 12 month outcomes were also assessed. However, not all subjects in the primary cohort were able to complete 12 month outcomes for a variety of reasons (including loss to follow-up, refusal to complete tests or physical inability). For this reason, a sub-set of 29 subjects statistically representative of the full cohort in terms of clinical and demographic characteristics (data not shown) was analysed for 12 month outcomes. Seven healthy male controls were also enrolled separately to determine serum hormone control values. Control subjects were enrolled if they did not have previous history of endocrinological pathology or previous brain injury.

Enrolled subjects with TBI were admitted to the neurotrauma intensive care unit to receive treatment consistent with The Guidelines for the Management of Severe Head Injury [34]. Temperature was monitored regularly and a small sub-set of subjects received moderate hypothermia (temperature 32.5–33.5°C for 48 hours) if they were enrolled in a randomized controlled clinical trial evaluating moderate hypothermia after severe TBI. Regardless of acute care study participation, subjects not receiving hypothermia were treated to maintain a normothermic state during their intensive care treatment. In total, six individuals in the cohort received hypothermia and 32 remained normothermic.

Serum sample collection

For enrolled subjects with TBI, blood was collected at ∼7:00 am daily for the first week when they were clinically available for the blood draw. Blood was then collected every 2 weeks until 6 months post-injury. Another sample was taken at 1 year post-injury. Upon collection and centrifugation, samples were aliquoted and stored at −80°C until the time of assay. For healthy control subjects, blood was similarly processed and stored for batch analysis. A total of 383 samples were measured in the TBI cohort. A total of 280 samples were collected between 2–52 weeks post-TBI. An additional 103 acute samples were collected on days 3–7 (hours 48–168) post-injury and were averaged to provide a value for each subject for week 1.

Serum TEST and E2 were measured in duplicate using radioimmunoassay with the Coat-A-Count® In-vitro Diagnostic Test Kit (Siemens Healthcare Diagnostics Inc., Los Angeles, CA). Each kit was a solid-phase 125-I radioimmunoassay designed for the direct, quantitative measurement of each hormone in serum; 25 µl (TEST) or 100 µl (E2) sample aliquots were evaluated in duplicate according to manufacturer instructions. The inter-assay and intra-assay coefficients of variation (CV) were less than 10% for these assays. FSH and LH levels were measured in duplicate using a highly sensitive fluoroimmunometric assay (Delfia, Perkin Elmer-Wallac, Turku, Finland). The inter-assay and intra-assay CVs were less than 10%. Samples with out of range (low) levels were assigned the detection limit of the respective assay.

Hypogonadotropic hypogonadism categorization

Normal serum hormone level ranges clinically used in determining the categorization of hypogonadotropic hypogonadism were taken from the University of Pittsburgh Medical Center's Presbyterian & Shadyside Automated Testing Laboratory [35]. To the authors’ knowledge, there are no published guidelines for establishing the diagnosis of hypogonadal hypogonadism using longitudinal data collected across a fixed period of time. Thus, when assigning hypogonadal status over time for these subjects, the clinical criteria described below were applied for each time point and then an arbitrary definition of hypogonadal hypogonadism was set to categorize hormone status for subjects during the first week and also during the span of weeks 1–52. Using this approach, subjects with two or more measurements on days 3–7 post-TBI were categorized a priori as AHH if 50% of these measurements had testosterone values <10 nmol L⁻¹ (minimum normal level) and LH values <5.6 IU L⁻¹ (maximum normal level). Also, each subject was a priori categorized as having PHH if at least 50% of all measurements, including both acute and 2–52 weeks post-TBI, had testosterone and LH values that fell in the ranges described above. Subjects having less than 50% of time points with values consistent with a clinical presentation of hypogonadotropic hypogonadism during this same time period were classified as having ERHH. Of the 38 subjects, n = 14 were classified as having PHH, with the other n = 24 classified as having ERHH. Hormone levels for the ERHH and PHH groups were compared to hormone levels measured in healthy controls (HC) (n = 7). Both hypogonadism status and hormone levels over time were compared to outcome at 6 and 12 months post-TBI.

Demographic and injury data

Demographic and pre-morbid variables including age, body mass index (BMI) and education (in years) were recorded for the TBI cohort. Mechanism of injury and hypothermia status were also recorded. Injury severity information was collected for each subject after resuscitation and without the influence of paralytics. This included injury severity scores (ISS), which incorporate injuries sustained to the whole body and initial hospital GCS scores, which specifically measure neurological injury. The ISS is derived from AIS (Abbreviated Injury Scale) scores, with a total score calculated by adding the sum of the squared AIS values assigned to the three most injured of six body regions [36]. The GCS is commonly used in studies to determine severity of neurologic injury after trauma and has been used in conjunction with other variables as a predictor of mortality and gross outcome [37]. Type of injury and head CT findings, based on admission head CT, were recorded and described. Radiographic evidence of any hypothalamic- or pituitary-specific injuries was drawn from all available head CT and MRI radiological reports in the patients’ medical records. Medications were reviewed at the time of enrolment and medication information was recorded at 6 and 12 months post-injury when available to screen for use of exclusionary medications.

Functional outcome measures

At 6 and 12 months post-injury, individuals were administered Glasgow Outcome Scale (GOS) scores. The GOS is a widely utilized measure that classifies outcome into five categories: 5 = good recovery, 4 = moderate disability, 3 = severe disability, 2 = persistent vegetative state and 1 = death [38]. Subjects were also administered the Disability Rating Scale (DRS) at 6 months post-injury, which rates individuals on four categories: arousal and awareness, cognitive ability to complete self-care functions, physical dependence on others, and psychosocial adaptability for work, housework and school. Total scores range from 0–30, with higher scores indicating a poorer outcome [39].

The Functional Independence Measure (FIM) was also assessed at 6 and 12 months post-injury by qualified, trained interviewers in accordance with Traumatic Brain Injury Model Systems criteria. This study specifically analysed the cognitive sub-scale, labelled FIM-COG, which represents a sum of scores for five items on the FIM pertaining to comprehension, expression, social interaction, problem-solving and memory. Each question is rated on a 7-point scale and summed to make up a total FIM-COG score ranging from 5–35, with higher scores indicating greater independence. This FIM-COG sub-scale has been used in other (TBI) studies as a measure of cognitive functional outcome [40], [41].

Cognition composite measure

A cognition composite measure was developed using a battery of neuropsychological measures that evaluate attention, episodic memory, language fluency and executive function, all of which have been previously utilized in the TBI population. Past research has shown the development of a composite measure as a practical means of evaluating general cognitive function in terms of recovery and performance, with increased consistency attained by aggregating multiple tests [42–45]. The composite measure of cognitive function was developed based on prior composite measures reported in the TBI population [46–48].

For this study, measures that were incorporated into the cognition composite measure were the Rey-Osterrieth Complex Figure, WAIS-R Digit Span, Wisconsin Card Sorting Task, Controlled Oral Word Association Test (COWAT) animal category, Trails A & B, Selective Reminding Test, Stroop Neuropyschological Screening Test and Symbol Digit Modalities Test. The Rey-Osterrieth Complex Figure is a visuospatial construction and memory task that requires that the subject copy a complex figure, reproduce the figure from memory immediately following completion of the copying task and reproduce the figure again 20 minutes later. Scores are based on the number of features correctly reproduced using standardized criteria [49]. Digit Span, a sub-test from the Wechsler Adult Intelligence Scale-R, requires individuals to repeat a sequence of numbers both forward and backward as a measure of memory and attention [50]. On the Wisconsin Card Sorting Task, an executive control task that assesses concept formation and cognitive flexibility, a subject must sort response cards according to specific performance constructs that require the ability to generate, switch and maintain response strategies. Performance is based on the ‘number of sorts’ achieved, total number of correct card placements and number of perseverative errors [51]. The Controlled Oral Word Association Test is used for assessing phonemic verbal fluency, i.e. the ease with which a person can think of words that begin with a specific letter, and semantic fluency, i.e. generating words that belong to a specific category (animals was the category used) within 1 minute [52]. The Trail Making Test parts A & B measures attention, visual searching and the ability to mentally control simultaneous stimulus patterns. Part A requires a subject to consecutively connect circles numbered in successive order from 1–25, while Part B requires a subject to consecutively connect circles while alternating between numbers (1–13) and letters (A–L). Performance for both tasks is measured using time (in seconds) to complete each part [53]. The Selective Reminding Test is a list-learning task in which the subject is asked to recall as many words as possible and is reminded of words omitted in that trial, until all words are learned or until the number of trials reaches a criterion [54]. The Stroop Neuropsychological Screening Test measures selective attention and cognitive flexibility by testing an individual's ability to meet changing cognitive demands and to suppress habitual responses in favour of more effortful ones [55]. The Symbol Digit Modalities Test involves a simple substitution task in which the examinee has 90 seconds to pair specific numbers with given geometric designs with written or oral responses. It is a measure of complex attention and psychomotor speed [56], [57].

Similar to other published work [45], the composite measure was created by converting raw scores for each individual, generated from each cognitive test variable, to a rank score with lower rank indicating a worse score on the test and a higher rank indicating a better score on the test. The ranked scores were then summed to create a composite score in which a lower composite signifies worse cognitive function and a higher composite signifies a better cognitive function. Because the composite measure uses rank comparisons, the scores are population-specific and based on the number of subjects in the cohort and the number of tests provided. Thus, the 6- and 12-month scores are markedly different in terms of the range of scoring.

Statistical analysis

Descriptive statistics, including mean and standard error of the mean, for continuous outcome variables and hormone levels were calculated and graphed separately for the PHH and ERHH groups. Week 1 hormone levels were averaged from the daily levels of days 3–7 post-injury, excluding days 1–2 in order to accommodate for the approximate time course for the development of acute gonadotropin and gonadal hormone deficits in the acute phase documented in previous studies [14]. Outlier hormone level values outside of 3 SD of mean hormone levels were removed from the data set prior to analysis. Average acute hormone levels for each group were graphed using mean values obtained from days 3–7 (labelled week 1). Subsequent values were obtained and averaged from samples collected bi-weekly for the first 6 months and at 1 year. To summarize, PHH associations with specific hormone levels, graphs were also made for the same time points, with data from weeks 2–24 averaged into one value. Healthy control values and clinical lab normal ranges were included on each graph for reference. This study examined bivariate associations between demographic/clinical variables and outcome variables in order to identify variables that should be controlled for in the multivariate models. A separate multivariate model was then built for each outcome variable (GOS, DRS, FIM-COG and Composite score) in order to assess the effect of hypogonadism on outcome, controlling for clinical/demographic variables. For the purpose of this study, GOS was collapsed into two groups, vegetative/severe (2–3) and moderate/good (4–5) and the effect of hypogonadism on this outcome measure was assessed using logistic regression. Because the remaining outcome measures (DRS, FIM-COG, Composite scores) were not normally distributed, non-parametric regression was used, specifically generalized semi-parametric additive models to examine the effect of hypogonadism on each of these outcomes while controlling for potential confounders identified in the bivariate analysis. Variables selected for multivariate testing had a p-value of <0.2 based on bivariate comparisons. When creating final models, only those variables with a p-value < 0.2 were kept in the model. All statistical tests were 2-tailed and multiple testing was corrected for using the False Discovery Rate method [58]. After correcting for multiple testing, the new cut-off for statistical significance was set at α = 0.044. Average hormone levels over the first week and from weeks 2–52 were correlated to outcome performance at 6 and 12 months using Spearman's and Pearons's correlations as appropriate. Statistical analyses were performed using SPSS Version 17.0 (Chicago, IL) and SAS Version 9.2 (Cary, NC).

Results

Description of the population

Table I summarizes the demographic and injury information for the entire cohort and the two groups based on hypogonadism status. The entire cohort of n = 38 had a mean age of 30.3 ± 1.9 years, mean BMI of 26.4 ± 0.6 and mean education of 12.7 ± 0.3 years. For injury severity, the mean ISS was 34.1 ± 1.7 and median GCS was 7. All six subjects who received hypothermia were in the ERHH group; 100% of the population met criteria for AHH and n = 14 were categorized as PHH, while n = 24 were categorized as ERHH. There were no significant relationships between hypogonadism status and any of the clinical variables, except hypothermia treatment and diffuse axonal injury (DAI). However, with the exception of composite scores at 12 months and hypothermia, neither hypothermia treatment nor DAI significantly impacted outcomes. Only one subject showed radiographic evidence of a hypothalamic injury, caused by a left frontal EVD catheter tip and noted on day 5 after TBI. Upon review of available CT images for each patient, none showed anatomical evidence of pituitary-specific injuries.

Table I. .

Demographics

Clinical and demographic information
Hypogonadism status	n	Hypothermia received (subjects)		Age Mean ± SE	BMI Mean ± SE	Education (years) Mean ± SE	ISS Mean ± SE	GCS Median
all	38	6		30.29 ± 1.87	26.42 ± 0.62	12.68 ± 0.34	34.05 ± 1.68	7
ERHH	24	6		30.38 ± 2.30	26.78 ± 0.83	12.83 ± 0.46	35.22 ± 2.17	6.5
PHH	14	0		30.14 ± 3.30	25.82 ± 0.92	12.39 ± 0.47	32.14 ± 2.64	7
p-value	n/a	<0.001		0.863	0.486	0.496	0.271	0.350
Mechanism of injury
		automobile		motorcycle	off-road vehicle	fall/jump	other	bicycle
all		14 (36%)		13 (34%)	4 (11%)	4 (11%)	2 (5%)	1 (3%)
ERHH		9 (38%)		7 (29%)	2 (8%)	3 (13%)	2 (8%)	1 (4%)
PHH		5 (36%)		6 (43%)	2 (14%)	1 (7%)	0 (0%)	0 (0%)
Chi-square p-value								0.880
Radiological injury type
	Subdural haematoma		Diffuse axonal injury	Epidural haematoma	Subarachnoid haemorrhage	Contusion	Intraventricular haemorrhage	Intracerebral haemorrhage
all	24 (63%)		10 (26%)	7 (18%)	24 (63%)	14 (37%)	7 (18%)	14 (37%)
ERHH	15 (63%)		10 (42%)	25 (86%)	7 (71%)	11 (46%)	5 (21%)	7 (29%)
PHH	9 (64%)		0 (0%)	1 (7%)	7 (50%)	3 (21%)	2 (14%)	7 (50%)
Chi-square p-value	0.912		0.006	0.227	0.201	0.175	1.000	0.201

Serum hormone associations with hypogonadism

After grouping subjects into the PHH or ERHH group based on the a priori categorization criteria for each post-injury sample collected as described above, longitudinal hormone profiles for the PHH and ERHH groups were plotted. Figure 1 shows group differences in aggregated data over different time periods during the first year post-TBI, while Figure 2 shows the full time course of serum hormone levels for subjects with PHH vs ERHH for TEST, E2, E2/TEST, LH and FSH compared to both healthy controls (HC) and clinical laboratory normal ranges.

Figure 1.

Differences in mean hormone levels–grouped time points by hypogonadal status. Grouped mean serum hormone levels by hypogonadism status over the first year after severe TBI compared to healthy controls (HC) and normal clinical laboratory ranges for (a) Serum testosterone levels. Testosterone levels are lower for PHH than ERHH at week 1 post-injury and other time intervals (p < 0.015 all comparisons). (b) Serum LH levels. LH levels are lower for PHH during week 1 and the 2–24 week interval (p = 0.046, week 1; p = 0.012, weeks 2–24). Both groups then reach HC levels at 6 and 12 months. (c) Serum estradiol levels. Estradiol levels are elevated in the first week for both PHH and ERHH groups. Estradiol levels for PHH are lower than ERHH for the week 2–24 time interval and at week 26 (p = 0.011, weeks 2–24; p = 0.019, week 26). (d) Serum estradiol/testosterone ratios. Estradiol/testosterone ratios for the PHH group are higher than the ERHH group during week 1, but are lower during the 2–24 week interval (p = 0.023, week 1; p = 0.011, weeks 2–24). (e) Serum FSH levels. FSH levels are lower for PHH during week 1 and weeks 2–24 (p = 0.038, week 1; p = 0.012, weeks 2–24). PHH, persistent hypogonadotropic hypogonadism; ERHH, early resolving hypogonadotropic hypogonadism; LH, Leutinizing hormone; FSH, Follicular stimulating hormone.

Figure 2.

Differences in mean hormone time course by hypogonadal status. Weekly mean hormone levels by hypogonadism status over the first year after severe TBI compared to healthy controls (HC) and normal clinical laboratory ranges for (a) Serum testosterone levels. Testosterone levels are lower in the PHH compared to ERHH levels in week 1. By 4–6 weeks post-TBI PHH testosterone levels remain low while ERHH testosterone levels gradually return to baseline. (b) Serum LH levels. LH levels are intermittently lower for the PHH group compared to the ERHH group. PHH group levels are lower than HC values for almost all time points except at 6 and 12 months. (c) Serum estradiol levels. Estradiol levels for the PHH group are lower than the ERHH group throughout the sampling period and never reach HC levels within 1 year. (d) Serum estradiol/testosterone ratios. Estradiol/testosterone ratios were elevated above HC's for both groups during week 1. The PHH group also had higher ratios again in the mid-chronic period. (e) Serum FSH levels. FSH levels were intermittently lower for the PHH group compared to the ERHH group during the first 12 weeks post-injury. PHH, persistent hypogonadotropic hypogonadism; ERHH, early resolving hypogonadotropic hypogonadism; LH, Leutinizing hormone; FSH, Follicular stimulating hormone.

Acute hormone associations with hypogonadism

As expected, TEST levels (Figures 1(a) and 2(a)) for the PHH group were significantly lower than the ERHH group during the first week post-injury (p = 0.015). LH for the PHH group remained below the ERHH group for week 1 (p = 0.046) (Figures 1(b) and 2(b)). Although E2 levels were not significantly different at week 1 by hypogonadism status (Figures 1(c) and 2(c)), E2/TEST ratios were elevated during week 1 for both groups (Figures 1(d) and 2(d)). However, PHH group ratios remained significantly higher than the ERHH ratios for the first week (p = 0.023). FSH levels (Figures 1(e) and 2(e)) were also lower for the PHH during the first week (p = 0.038, week 1).

Chronic hormone associations with hypogonadism

TEST levels (Figures 1(a) and 2(a)) for the PHH group were consistently and significantly lower than the ERHH group at later time intervals (p < 0.001, weeks 2–24; p = 0.010, week 26; p = 0.008, week 52; Figure 1(a)). TEST levels for the PHH group remained below HC values across the time course, while TEST levels for the ERHH group reached the bottom-range for normal clinical values by 4–6 weeks after injury and remained close to HC levels (see Table II for HC levels) from week 8 forward (Figure 2(a)). LH levels (Figures 1(b) and 2(b)) were also lower in the PHH group over time, although less consistently than that observed for TEST. LH for the PHH group remained below the ERHH group at weeks 2–24 (p = 0.012). E2 levels for the ERHH group reached the bottom-range for normal clinical values at ∼ 4 weeks after injury, but did not approach HC levels (see Table III) until weeks 22–24 (Figure 2(c)). However, mean E2 levels for the PHH group never reached the HC levels within the 1-year monitoring period. E2 levels were significantly lower for the PHH group compared to the ERHH group (p = 0.011, weeks 2–24; p = 0.019, week 26; Figure 1(c)). For E2/TEST ratios, both groups were significantly decreased within the first 4 weeks (Figure 2(d)) and approaching control levels. On average, PHH group ratios remained significantly higher than the ERHH ratios between 2–24 weeks (p = 0.011; Figure 1(d)), although the groups converge at weeks 4–8 and again at later time points nearing 6 months (Figure 2(d)). FSH levels (Figures 1(e) and 2(e)) were also lower for the PHH group over time, although less consistently than that observed for TEST and E2, remaining significantly below the ERHH group for weeks 2–24 (FSH: p = 0.012, weeks 2–24) and below HC values for almost all time points up to 6 months after injury.

Table II. .

Control serum hormone ranges

	Testosterone (nmol L−1)	Oestrogen (pg mL−1)	LH (IU L−1)	FSH (IU L−1)
UPMC pathology lab ranges	10.00–42.00	20.0–75.0	1.0–5.6	1.5–14.3
Wagner lab controls	15.88414 ± 7.375	37.296 ± 14.886	4.559 ± 2.473	5.075 ± 3.837

Table III. .

Clinical variable associations with outcome

Independent variable	GOS		DRS		FIM-COG		Composite score
	Beta	p-value	Beta	p-value	Beta	p-value	Beta	p-value
6 M (n = 38)
Intercept	− 17.12	0.008	20.2	<0.001	−1.73	0.852	87.08	0.567
Hypogonadism status	3.19	0.014	−3.66	0.018	5.78	0.022	163.46	0.017
Age					−0.16	0.133
GCS	0.53	0.128	−1.96	<0.001	2.60	0.003	49.93	0.028
BMI	0.43	0.022			0.51	0.112
12 M (n = 29)
Intercept	−4.10	0.082	19.78	<0.001	4.35	0.464	131.29	0.356
Hypogonadism status	1.55	0.140	−4.59	0.031	6.79	0.020	135.95	0.047
GCS	0.67	0.057	−1.92	0.007	3.10	0.002	31.61	0.146

Outcomes based on hypogonadism status

Figures 3 and 4 summarize univariate associations between hypogonadism status and outcome at 6 and 12 months after TBI. After adjusting for these co-variables in the multivariate models, it was found that subjects with PHH had a significantly greater proportion of lower (worse) GOS scores at 6 months (p = 0.014) when compared to subjects with ERHH and there was a trend for worse GOS scores at 12 months (p = 0.140). Similarly, the average DRS score for the ERHH group was lower (better) than the average DRS score for the PHH group at 6 months (p = 0.018) and 12 months (p = 0.031), suggesting the same association between hypogonadism and more disability (Figures 3(b) and 4(b)). The FIM-COG scores (Figures 3(c) and 4(c)) were significantly lower (worse) for the PHH group when compared to the ERHH group at both 6 months (p = 0.022) and 12 months (p = 0.020) after TBI. Cognitive composite scores (Figures 3(d) and 4(d)) were significantly lower (worse) in the PHH group when compared to the ERHH group at 6 months (p = 0.017), with a trend observed (after correction for multiple testing) at 12 months (p = 0.047) after injury. Table III shows demographic/clinical variables that were controlled for in the multivariate models.

Figure 3.

Hypogonadism associations with 6-month outcome post-injury. Bivariate association between hypogonadism and 6 month outcome. PHH was associated with worse 6 month (a) Glasgow Outcome Scale (GOS) scores (p = 0.016), (b) Disability Rating Scale (DRS) scores (p = 0.017), (c) Functional Independent Measure Cognition (FIM-COG) scores (p = 0.005) and (d) cognitive composite (p = 0.012) scores compared to the ERHH group. After potential confounders are controlled for using multivariate non-parametric regression, these relationships still hold (see Table III). PHH, persistent hypogonadotropic hypogonadism; ERHH, early resolving hypogonadotropic hypogonadism.

Figure 4.

Hypogonadism associations with 12-month outcome post-injury. Bivariate association between hypogonadism and 12 month outcome. (a) Glasgow Outcome Scale (GOS) scores were not significantly different for PHH vs ERHH (p = 0.18). PHH was associated with worse 12 month (b) Disability Rating Scale (DRS) scores (p = 0.027), (c) Functional Independent Measure Cognition (FIM-COG) scores (p = 0.014) and (d) cognitive composite scores (p = 0.026) compared to the ERHH group. These relationships still hold after potential confounders are taken into account in the multivariate models (see Table III). PHH, persistent hypogonadotropic hypogonadism; ERHH, early resolving hypogonadotropic hypogonadism.

Hormone levels and outcomes correlations

TEST and E2 levels obtained weeks 2–52 post-TBI were significantly correlated with all outcomes. Lower TEST and E2 levels were associated with worse outcomes at 6 months (TEST: │r = 0.35–0.537│; p ≤ 0.015, E2: │r = 0.291–0.509│; p ≤ 0.041 all comparisons). There were similar associations for all 12 month outcomes (TEST: │r = 0.334–0.439│; p ≤ 0.027, E2: │r = 0.278–0.401│; p ≤ 0.050 all comparisons) with the exception of DRS scores and TEST (p = 0.068). Interestingly, there were trends for acute (week 1) TEST levels to be associated with 6 and 12 month composite scores (p ≤ 0.052 both comparisons), where lower TEST was associated with worse outcome. While there were no significant correlations between acute E2 levels and any of the outcomes, lower acute E2/TEST ratios during the first week were associated with better outcomes. Lower acute E2/TEST ratios were correlated with better (lower) 6 month DRS (p = 0.044) and better composite scores (p = 0.034) as well as better 12 month GOS (p = 0.050) and composite scores (p = 0.014), with other comparisons showing similar trends. However, E2/TEST ratios measured 2–52 week post-injury were not associated with outcome. Interestingly, higher LH values during the first week post-injury were significantly correlated with better outcome for 6 and 12 month composite scores (6 month: p = 0.010; 12 month: p = 0.048), 12 month GOS (p = 0.049) and 12 month FIM-COG (p = 0.028), with similar trends observed for other outcomes measures.

Discussion

The hormonal response after TBI is complex and varies over time. While the acute effects of TBI on serum hormone levels have been previously examined, the chronic effects and how they distinctly differ from the initial hormonal response, have not been fully explored [6], [7]. This work suggests a relatively novel concept within the post-traumatic hypogonadism literature. Prior to examining hormone levels within this population, clinical criteria were adapted for the longitudinal assessment of hypogonadism over time. It was found that, while all individuals are affected by hypogonadotropic hypogonadism acutely, there appears to be a distinct time window and pattern whereby the hormonal deficits resolve for some but not others. The results of this study describe the varied sex hormone responses to TBI and specifically describe a transition from AHH into either ERHH or PHH. Some previous work similarly describes a pattern of acute dysfunction and chronic dysfunction vs recovery [59]. Further, the work presented in this study is important in that it suggests novel relationships between hypogonadism and outcome post-TBI. Accordingly, this detailed timeline and hormone characterization for the onset and persistence of hypogonadism post-TBI may lead to better screening and management that could improve outcomes for this population.

Examination of the acute and post-acute course of hormones after TBI suggested that, although PHH TEST levels are lower than the ERHH levels acutely, they converge at weeks 2–4 post-injury. Beyond this point, hormone levels for the ERHH group gradually returned to normal levels while the PHH group levels remained below the normal range noted for the clinical laboratory, reflecting the associations between low TEST and worse long-term outcome. While E2 levels for both groups were relatively high acutely, E2 precipitously decreased to deficient levels by week 2 and the levels for both groups converged at this point in the early post-acute phase. E2 gradually increased to levels that are within the normal range for the clinical laboratory, although the PHH group remained significantly lower and lagged behind the ERHH group. This finding is also linked to the association between chronically low E2 levels and worse outcome. E2/TEST ratios were notably higher than controls for both groups acutely, particularly for the PHH group, which, based on previous work [14], may indicate an increased injury and/or stress response for this group. Interestingly, higher E2/TEST ratios during the first week post-TBI were associated with worse outcomes.

The association between better outcomes, higher acute testosterone levels and lower acute E2/TEST ratios seems to contradict, in part, previous studies suggesting that higher acute E2 and TEST are associated with poorer outcome. However, it is important to emphasize the distinction that this current study's acute assessment includes only survivors and hormone level measurements started at day 3 (instead of day 1) in order to assess people at or after the onset of acute hypogonadism [14]. As suggested in other studies, the pituitary shuts down quickly and acute hormonal physiology after severe TBI primarily is driven by adrenal activity and/or peripheral aromatase activation that leads to relatively elevated E2 levels for those with worse outcomes [11–13]. For those with PHH, E2/TEST ratios remained higher (i.e. higher aromatization) for week 1 and weeks 2–24 post-injury, although levels for weeks 2–24 were no longer directly associated with outcome in this small cohort. However, this data does suggest that the chronically higher hormone ratios associated with the PHH group may represent a possible ongoing stress or inflammation state [60], [61], where aromatase is more active, even in the setting of low testosterone and E2 levels. Despite higher E2/TEST ratios, E2 levels in the PHH group were low and associated with worse outcomes. Although BMI was associated with 6 month GOS (p = 0.022) and may indirectly represent peripheral aromatization, in this small cohort, this variable was not associated with hormones in the acute or post-acute phase. Better outcomes and ERHH status associated with higher LH levels during week 1 may be an indicator of subsequently preserved pituitary function over the recovery time course assessed in this study.

E2 has a positive role in neuroplasticity and is associated with the presence of other neuroplasticity modulating molecules like neurotrophins [62–64]. It also has a role in modulating GABAergic and glutamatergic tone [65] and outgrowth of synaptic processes [66], [67] in order to promote a neuroplastic environment. An increasing number of studies in the literature suggest that E2 is an important factor for memory and cognition [68]. The relative lack of E2 in the PHH group, compared to ERHH, may be one reason for worse recovery in this group [24]. While TEST appears to modulate neurogenesis and neuroplasticity in some animal models [69–71], other studies suggest that higher E2 levels are associated with cognitive improvements in men receiving testosterone treatment for idiopathic hypogonadism [20]. Further, endogenous TEST and E2 levels are linked to neuropsychological testing performance in healthy ageing men, some of whom met criteria for idiopathic hypogonadism [72]. Thus, it is possible that both E2 and TEST levels had an impact on cognitive and outcome measures in the population with post-traumatic hypogonadism. The significant associations in this study between hypogonadism, low E2 and TEST levels and worse outcome highlight the potentially negative consequences of TBI-mediated hypogonadism on outcome and recovery. However, routine screening for hypogonadism after TBI should be considered because the consequences of hypogonadism may not be overtly apparent in terms of clinical presentation and symptoms, especially since clinical signs and symptoms may overlap with deficits secondary to the TBI. Further work is required to identify and understand the effects of TBI on hypogonadism in relation to depression, quality-of-life and other symptoms, such as fatigue. In some studies, hormone replacement helps reverse symptoms associated with depression in other non-TBI populations [73–75]. Importantly, the necessity for routine screening is further supported by the relative lack of correlations between demographic/clinical variables and hypogonadism or outcome, which makes it harder to predict who is at risk for PHH. Some literature suggests that hypogonadism is related to injury severity [59]. Among this population, GCS was not associated with the development of PHH, which is likely due to the limited GCS range in a small cohort that is restricted to severe TBI.

There was a significant relationship between hypogonadism status, hypothermia treatment and DAI, as all subjects who received hypothermia and/or had DAI were in the ERHH group (50% of the hypothermia patients also had DAI). This may suggest hypothermia and/or injury type (DAI) to be indirect factors with hypogonadism status and outcome. The study is not adequately powered to fully examine relationships between these co-variables, hypogonadism and outcome. Also, a full characterization of DAI severity, known to impact outcome [76], was not undertaken. While several animal studies have successfully established the protective use of hypothermia [77], [78], clinical trials have had less consistent results [79], with one multi-centre trial concluding that hypothermia treatment was not effective in improving global outcomes after severe TBI [80]. However, the primary purpose of this study was to examine the fundamental relationship between hypogonadism post-TBI and outcome, regardless of treatments or other injury characteristics that may influence the resilience of the HPG axis and hormone profiles. Nonetheless, neuroanatomic and radiographic associations with hypogonadism, as well as potential benefits of hypothermia on post-traumatic HPG function could be interesting areas of additional study.

Contrary to other literature suggesting only a limited value of early hormones in predicting later hypogonadism [81], average acute hormone levels for TEST, E2/TEST, FSH and LH in this study were lower for the PHH vs the ERHH group. This result suggests that early (week 1) hormone levels may be helpful in predicting hypogonadism and generating prediction rules for the development of persistent hypogonadism. The study results also suggest that routine screening should occur early, at ∼2 weeks post-injury and again at 6–8 weeks post-injury, when hormone levels either remain low or begin to normalize. These proposed guidelines contrast with prior studies that have assessed chronic hypogonadism following TBI by examining testosterone levels at longer intervals (3, 6 or 12 months) post-injury and suggesting that testing for hypogonadism be repeated again no sooner than 1 year following injury [18], [82–84]. However, these recommendations are not based on the concept that AHH transitions into either PHH or ERHH or the timing in which hypogonadism either persists or resolves. Based on these results, recommendations for follow-up testing at least 1 year out from injury may preclude the possibility of hormone replacement mediated improvements in outcome at 6 and 12 months. Future work should include long-term hormone testing to determine if hypogonadism resolves after the first year of recovery for those with PHH or if it persists beyond 1 year and/or can be considered a permanent condition. In addition, it may be beneficial to examine hormone levels at multiple points over a greater span of time for those who have recovered early from post-traumatic hypogonadism (ERHH group). Most men have a decline in their testosterone levels with increasing age. Thus, in men with TBI whose hypogonadism initially resolves, it may be possible that they later experience age-related declines in hormone levels that are not typical in the uninjured population and may contribute (along with the TBI) to more prominent age-related declines in cognition and physical function that are known to occur in the population with idiopathic hypogonadotropic hypogonadism [19].

In conjunction with the adverse effects of hypogonadism on cognitive recovery from TBI, undiagnosed hypogonadism may have other relevant consequences that are further exacerbated by the long durations of immobility that often accompany severe TBI. These problems include fatigue, muscle weakness, reduced lean body mass, reduced bone mass and impaired exercise tolerance, related to testosterone deficiency in men [85]. Ongoing hypogonadism is associated with diabetes, metabolic syndrome and increased risk for cardiovascular disease [86]. Sex-hormone replacement for hypogonadism due to various aetiologies is well-documented as having beneficial effects, such as increased vigour, muscle strength, self-esteem and libido as well as reduced fatigue [85], [86]. However, the benefits of sex-hormone replacement specifically for TBI-mediated hypogonadism have not been adequately studied yet. Hormone replacement is not without risk with regard to testicular and prostate pathology [86], [87]. The risk/benefits of hormone replacement post-TBI are not clear, but it seems reasonable to speculate that appropriate testosterone replacement in the post-TBI setting could aid in both cognitive and physical recovery. To date, there is no consensus on how and when to treat post-traumatic hypogonadism [59]. However, when considering treatment for post-traumatic hypogonadism, it is important to remember that TBI can result in a number of neuroendocrine abnormalities that involve the hypothalamic-pituitary-adrenal (HPA) as well as the hypothalamic-pituitary-gonadal (HPG) axes [84]. Deficits in cortisol, thyroid and growth hormone can present concurrently with hypogonadism and can have overlapping symptom profiles. When considering the possibility of hormone replacement for hypogonadism, a thorough work-up of all potential hypothalamic pituitary deficits should occur prior to treatment. First cortisol and then thyroid replacement in those with deficits is required prior to androgen replacement in order for the body to adequately respond to the increases in metabolism that accompany testosterone therapy [88]. Initial correction of cortisol and/or thyroid deficits may resolve many of the overlapping symptoms of these pituitary endocrinopathies such that decisions on whether to treat residual gonadal hormone deficits can be specifically addressed. Growth hormone should also be assessed as it is often abnormal in the setting of other acquired hormone deficiencies [88].

Although some individuals sustain physical damage to anatomic structures, such as the pituitary and hypothalamus, which contribute to hormonal dysfunction, even those with mild injuries may be susceptible to developing hypogonadism [59]. The data based on head CT show that anatomic evidence of direct damage to hypothalamic/pituitary structures is rare even after severe TBI. More subtle injuries, including secondary hypoxemia, can affect the pituitary without evidence of physical damage [18], [82]. Also anti-pituitary antibodies may play a role in the development of persistent or permanent post-traumatic hypogonadism [89]. Experimental evidence suggests that neuroinflammation and reactive gliosis occur in pituitary tissue after experimental TBI [90]. As such, it may be reasonable to speculate that persistent neuro-inflammation after TBI could lead to conditions similar to autoimmune hypophysitis in which significant lymphocytosis leads to pituitary stalk thickening and the progressive development of panhypopituitary deficits [91], [92], a condition most commonly associated with hypocortisolism.

In contrast, other sources for hypopituitarism should be examined. For example, while hypocortisolism is a fairly frequent presentation with post-traumatic hypopituitarism [22], [82], [83], [93], hypercortisolism can diminish gonadal hormone production in chronic stress paradigms [94], [95]. Therefore, cortisol influences on PHH may be relevant after the stress of major trauma associated with severe TBI and the development of TBI-related conditions, like post-traumatic depression. Genetic variability for the enzymes synthesizing these hormones and for sex hormone binding globulin may also contribute to hormone profiles and outcome. For example, some of the initial work implicates the aromatase gene in early CSF hormone levels and TBI outcome [96]. The range of potential mechanisms underlying the development of PHH warrants future studies specifically evaluating the causative factors underlying this common problem after TBI.

Although gender studies have been mixed when assessing how gender may influence TBI pathology, recovery and outcome [14], [20], [97–102], hormone research after TBI has been focused on gonadal hormone-mediated neuroprotection. Previous pre-clinical studies demonstrate the possible benefits of estradiol [100] and progesterone [103], [104] in conferring neuroprotection. Also, clinical studies suggest that pharmacological treatment doses of progesterone are neuroprotective [105], [106]. In contrast, previous work shows that endogenous serum hormone profiles in those with severe TBI are not neuroprotective, but, rather, represent an acute stress response to critical illness that may also adversely impact outcome. While acute CSF hormone profiles may have a different role in TBI pathology, the work here also shows that chronic endogenous serum hormone profiles may have a unique influence on TBI outcome, specifically one where hormones like estradiol may have beneficial effects on neuroplastic mechanisms underlying recovery.

Limitations

The primary limitation with this study is the small cohort size. Examination of a larger cohort may provide more generalizable results. An additional issue with this cohort is the limitation of only including subjects who completed all evaluated outcomes or who were cognitively unable to complete some neuropsychological measurements. This excluded some potential subjects who refused or were physically unable to complete tests, as well as anyone lost to follow-up, each of which could have led to selection bias for this population. Another limitation is that only raw data were used for the neuropsychological measures and, therefore, were not adjusted for age and education. However, demographic variables that were found to have an effect on outcome were controlled for in the multivariate model. When making the clinical diagnosis for hypogonadism in men, many clinical laboratories measure total (protein bound) serum TEST concentration, but some also estimate free or bioavailable TEST, either by direct measurement or indirectly by measuring sex-hormone binding globulin (SHBG) level and calculating the free TEST index (total TEST/SHBG ratio) [107], [108]. The role of free vs total hormone levels has been considered with other disorders and should be studied for its role in the diagnosis of post-traumatic hypogonadism, sympotomatology and associated potential treatment [109].

Conclusion

TBI is a critical problem associated with extensive disability and high morbidity. This study demonstrates that post-TBI pituitary dysfunction is relatively common and that it contributes to poor outcome and recovery. Considering the high incidence of TBI, resultant hypogonadotropic hypogonadism arises as a significant public health issue that, for most individuals, fails to be identified or treated. Hypogonadotropic hypogonadism leads to worse cognitive and global outcomes in the exploratory analysis of this male cohort with severe TBI, a finding that appears to be mediated by sex hormones. Because the condition is often overlooked, it may lead to further complications and difficult recovery. Appropriate screening for hypogonadism and further work investigating the mechanisms by which PHH occurs in order to define appropriate management strategies may optimize outcome and recovery.

Acknowledgements

The authors would like to acknowledge Mary Synnott for figure construction and the UPMC Trauma Registry team and the Brain Trauma Research Center for their role in data collection.

Declaration of Interest: This work was supported by CDC grant # R49/CCR323155 (AKW, SB) and NIH 5P01NS030318-16 to (CED). The authors report no conflicts of interest.