Hypothalamopituitary Dysfunction Following Traumatic Brain Injury and Aneurysmal Subarachnoid Hemorrhage:  A Systematic Review FREE

Traumatic brain injury (TBI) is a major public health problem with an overall incidence of 235 per 100 000 persons per year¹ and some 80 individuals per 100 000 population are hospitalized for TBI in the United States each year.² Traumatic brain injury is the leading cause of death and disability in young adults.³ Aneurysmal subarachnoid hemorrhage (SAH) occurs in 6 to 10 individuals per 100 000 population each year.⁴ Both TBI and SAH leave many survivors with significant adverse physical and psychological sequelae of the trauma.^5- 6

Hypopituitarism caused by TBI was first reported in 1918,⁷ only 4 years after the initial clinical description of hypopituitarism. However, in subsequent decades, TBI has been considered a rare cause of hypopituitarism. In 2000, a review reported 367 cases of posttraumatic hypopituitarism.⁸ Aneurysmal SAH has been considered to be an even rarer cause of hypopituitarism with almost no published data until recently.⁹

In the last few years, however, a number of systematic studies have shown hypopituitarism to be a common complication of both TBI and SAH.^10- 29 This adds an important new dimension to the current understanding of these conditions and raises questions about the need for pituitary screening for these patients.

The signs and symptoms of hypopituitarism may be subtle³⁰ and overlap with the neurological and psychiatric sequelae of head trauma and SAH (Table 1). This can explain why the diagnosis of hypopituitarism is often missed or delayed after these conditions—with potentially serious and sometimes life-threatening consequences for the affected patients.

This article reviews the research on hypothalamopituitary dysfunction as an underdiagnosed consequence of TBI and SAH, the natural history of this complication, and the potential clinical and public health implications of posttraumatic hypopituitarism.

We searched MEDLINE for articles published from 2000 to 2007. We used any combination of the terms traumatic brain injury or subarachnoid hemorrhage with that of pituitary, hypopituitarism, growth hormone deficiency, hypogonadism, hypocortisolism, hypothyroidism, or diabetes insipidus. In addition, we searched the reference lists of articles identified by this search strategy. We selected all articles reporting original data on endocrine outcomes after TBI or aneurysmal SAH with regard to prevalence, pathogenesis, risk factors, outcomes, and clinical course (Figure). Due to adaptive and functional endocrine changes in the acute phase that interfere with evaluation of pituitary function, we did not select articles reporting exclusively on endocrine findings in the acute phase.

We pooled data by adding all absolute numbers of patients included in the selected studies and patients with hypopituitarism (if available) and deficiencies of each pituitary axis. We calculated the percentages of hypopituitarism with 95% confidence intervals (CIs) in each single study and in the pooled data. We used SPSS version 13 (SPSS Inc, Chicago, Illinois) to calculate 95% CIs.

Case reports were cited only if no other data were available. Review articles were cited to give more general background information on the topic.

We found 19 clinical studies reporting on pituitary function after TBI or SAH, including a total of 1137 patients.^10- 29 Of these, 2 studies with 74 patients reported on pediatric populations,^24- 25 whereas the other 17 studies reported on adult populations. In some studies, the patients were tested more than once at different times while in other publications, data on both TBI and SAH are reported. For reasons of clarity, we counted reports in different publications on the same study populations (as far as evident from the publication) as 1 study and reports on TBI and SAH in a single publication as 2 studies.

Most studies were cross-sectional, whereas 5 studies were prospective and longitudinal with at least 2 endocrine evaluations at specified times. Tables 2 and 3 summarize design, patient selection, and characteristics of the various studies.

Regarding potential pathogenesis, we found 6 systematic studies evaluating pituitary pathology in 638 individuals with TBI^31- 35 and 2 studies evaluating hypothalamic damage in 106 individuals with TBI and 102 individuals with SAH.^34,36 Most of the pathological studies we found were published decades ago.

The anterior pituitary gland produces several peptide hormones that act on target organs peripherally: adrenocorticotropic hormone, thyroid-stimulating hormone, luteinizing hormone, follicle-stimulating hormone, prolactin, and growth hormone. Pituitary hormone secretion is regulated by hypothalamic-releasing hormones and inhibitory factors (which are released into the portal circulation in the pituitary stalk) and by negative feedback from peripheral hormones. Pituitary dysfunction may occur at the hypothalamic, stalk, or pituitary level.

Following TBI, the prevalence of endocrine dysfunction in all clinical studies ranged from 15% to 68% and the prevalence following SAH ranged from 37.5% to 55%. There was, however, considerable disagreement on the relative frequency of the various anterior pituitary axes affected.

Table 4 summarizes the results of 13 studies with 809 patients with TBI and 102 patients with aneurysmal SAH that were performed at least 5 months following the injury^{12,14- 23,27- 29} (we excluded studies in the early phase after injury^{10- 11,13,26} to avoid the confounding effect of acute critical illness on neuroendocrine function and studies on pediatric populations for reasons of homogeneity^24- 25). Of these, the pooled prevalences of anterior hypopituitarism after TBI and aneurysmal SAH were 27.5% (95% CI, 22.8%-28.9%) and 47% (95% CI 37.4%-56.8%), respectively. Aneurysmal SAH was associated with significantly higher frequencies of overall hypopituitarism, growth hormone deficiency, and corticotropic deficiency than TBI. After TBI, deficiencies of luteinizing hormone/follicle-stimulating hormone and growth hormone were significantly more common than adrenocorticotropic hormone deficiency, which was significantly more common than thyroid-stimulating hormone deficiency. Following SAH, deficiencies of growth hormone and adrenocorticotropic hormone secretion were more common than deficiencies of luteinizing hormone/follicle-stimulating hormone and thyroid-stimulating hormone.

When interpreting these results, however, it is necessary to consider that a laboratory value below a respective cut-off or threshold does not necessarily reflect clinically relevant impairment of the respective pituitary function. In general, normal endocrine values have been established in a healthy, mainly middle-aged, and normal-weight population. The patients with brain pathologies that were assessed in the studies included in this review often differ from these aspects. Also, the nonspecific health impairments of these patients might additionally influence hormone levels.

Thus, ideally, control groups without brain pathologies that have been matched for sex, age, body mass index, and severity of health impairment should have been included in these studies. This has not been the case, but in 4 studies, healthy control groups matched by sex, age, and body mass index have been included for the evaluation of at least some critical endocrine tests.^10,15,17,21 In the studies by Kelly et al¹⁰ and Popovic et al,¹⁷ the values obtained in the control groups were used to establish reference ranges for the studied patients with TBI. In the studies by Agha et al¹⁵ and Schneider et al,²¹ it was additionally indicated how many controls and patients had values below the respective threshold. For the evaluation of somatotropic function, 0 of 31 (0%) and 1 of 38 (2.6%) control individuals failed the respective stimulation tests in the studies by Agha et al¹⁵ and Schneider et al,²¹ whereas 18 of 102 (17.6%) and 7 of 77 (9.2%) patients with TBI failed the same tests. This indicates that abnormal hormone values are ascertained in patients with TBI with a much higher frequency than in the control population.

The posterior pituitary is a storage organ for the hypothalamic hormones oxytocin and antidiuretic hormone (or vasopressin). Following TBI or SAH, posterior hypopituitarism presents with central diabetes insipidus (with potentially life-threatening hypernatremia if the patient has impaired thirst or inadequate fluid intake). In a cross-sectional and a prospective-longitudinal study of posterior pituitary function after TBI, the prevalence of diabetes insipidus, diagnosed using the criterion standard water deprivation test, was 26% in the acute phase¹⁹ and 6.9% among long-term survivors.¹⁶

In the 40 patients studied by Kreitschmann-Andermahr et al²⁷ at least 1 year following aneurysmal SAH, no patient had clinical evidence of diabetes insipidus, whereas Aimaretti et al¹⁴ reported diabetes insipidus—assessed by diuresis, urine density, serum sodium levels, and plasma osmolality—in 2.8% of 32 SAH survivors 12 months after the event.

Untreated hypopituitarism is associated with serious morbidity^30,37- 39 and premature mortality.^40- 41 Serious and life-threatening adrenal crises secondary to acute adrenocorticotropic hormone deficiency in patients with TBI have been highlighted in the literature^42- 43 with dramatic improvement following glucocorticoid replacement.⁴² Increased neuropsychiatric morbidity in patients with TBI with growth hormone insufficiency also has been reported.⁴⁴ In a recent large study of 104 patients with TBI, posttraumatic hypopituitarism was independently associated with poor quality of life (particularly in scores of energy, sleep, and physical mobility), abnormal body composition, and adverse metabolic profile, a pattern of impairments also seen in patients with hypopituitarism of other etiologies.⁴⁵ In patients with SAH, preliminary data indicate that neuroendocrine disturbances contribute to disturbed quality of life, depression, and sleeping disturbances.⁴⁶ These findings indicate that hypopituitarism after both TBI and SAH is associated with poor recovery and worse outcome.

Considering that hypopituitarism appears to be a common occurrence following TBI and SAH, most cases likely remain unrecognized and untreated. Moreover, other studies suggest that other causes of acquired brain injury including ischemic stroke, cranial irradiation, or surgery for nonpituitary tumors may also be important but underappreciated risk factors for hypothalamopituitary dysfunction.^47- 49 Although some of the pituitary hormone deficiencies identified in the different studies are partial and could be of uncertain functional significance in an otherwise healthy individual, they may have added significance in patients with TBI both in the acute phase and during rehabilitation, thus justifying screening of patients with TBI and SAH who are at risk of this important but treatable complication.

The postresuscitation Glasgow Coma Scale (GCS) is the most widely used tool to assess the severity of brain injury in clinical practice. It assesses eye-opening, verbal, and motor responses on a scale ranging from 3 to 15. Brain injury severity is graded as severe (GCS score of 3-8), moderate (9-12), and mild (13-15).⁵⁰ Some studies included patients with all grades of TBI severity,^{12,14,21- 22} while others included patients with moderate or severe TBI^15,17 or only severe TBI.^18,23 In some studies no associations of hypopituitarism with severity of TBI were reported,^{14- 15,17,21- 23} whereas others found hypopituitarism to be more frequent in patients with more severe TBI.^12,29 We further analyzed all studies that included all grades of TBI in the chronic phase and reported the prevalence of hypopituitarism according to postrescuscitation GCS score.^{12,14,21- 22,29} The pooled prevalences of hypopituitarism with severe, moderate, and mild TBI (as defined by the GCS) were 35.3% (95% CI, 27.3%-44.2%), 10.9% (95% CI, 5.1%-21.8%), and 16.8% (95% CI, 10.9%-25.0%), respectively.

No association between hypopituitarism and cranial computed tomography (CT) imaging results was found in one study¹⁵ while a small study found an association between hypopituitarism and diffuse brain swelling on cranial CT.¹⁰

In a reanalysis of the study by Schneider et al,²¹ diffuse axonal injury and basal skull fracture were risk factors for posttraumatic hypopituitarism.⁵¹ Another study reported increased intracerebral pressure to predict posttraumatic hypopituitarism.²⁹ Herrmann et al²³ found an association between length of stay in the intensive care unit and hypopituitarism. In a retrospective analysis, abnormalities in pituitary magnetic resonance imaging or CT scans were found in 80% of patients with TBI with hypopituitarism compared with only 29% of patients without hypopituitarism.⁵² Posttraumatic diabetes insipidus was not associated with presence of anterior hypopituitarism but was associated with more severe head trauma and the presence of cerebral edema on CT scan.¹⁶

In patients with SAH, no association between clinical state on admission to the hospital as measured by the Hunt and Hess Scale (ranging from slight meningism to coma),⁵³ the amount of blood on the initial CT scan (Fisher CT score),⁵⁴ or any other clinical parameter known to be associated with a poor clinical outcome after SAH and risk of hypopituitarism was found.^14,27- 28

In summary, clinical severity of aneurysmal SAH does not help discriminate between patients at high and low risk of developing hypopituitarism but patients with severe TBI are at higher risk of hypopituitarism. However, even patients with mild TBI still have a lower but substantial risk of hypopituitarism. Thus, in general, all patients hospitalized for TBI or SAH should be considered for endocrine evaluation. However, special attention should be paid for severe TBI, basal skull fractures, diffuse axonal injuries, increased intracranial pressure, and prolonged stay in intensive care.

Table 5 summarizes the results of prospective studies analyzing pituitary function at different points after TBI. In 2 studies, the initial assessment was performed in the acute phase first,^19,22 whereas in the other 2 studies, the initial assessment was performed at 3 months after trauma.^14,21 All studies repeated testing 12 months after trauma. In 1 study, endocrine testing additionally was performed 6 months after TBI.¹⁹ In all studies, there was a trend toward improvement in pituitary function over time and some of the early abnormalities were transient with complete recovery. Conversely, hypopituitarism evolved over time in some patients and became detectable in the postacute phase or later during rehabilitation. In 1 study, further improvement of pituitary function occurred between 6 and 12 months after trauma.

New onset of deficiencies was observed between 0 and 6 months, but no new deficiencies occurred thereafter.¹⁹ In other studies, new deficiencies occurred only rarely (5%) after 3 months and these were only single-axis deficiencies. Only 1 study evaluated pituitary function 3 and 12 months after SAH. The rate of pituitary dysfunction also decreased from 47% to 38% by 1 year.¹⁴ No new deficiencies occurred in patients with normal pituitary function at 3 months after SAH.

The fact that brain injury–induced neuroendocrine dysfunction has attracted little attention for so many decades is, in retrospect, surprising because large neuropathological studies including a total of 638 cases have established a large frequency of 26.4% to 86% hypothalamic-pituitary damage in patients who died as a consequence of TBI.^31- 35,55 Autopsy results have demonstrated different types of lesions from damage to the pituitary capsule (the most frequent form of pituitary damage after TBI, occurring in 23.3%-59%) to injury to the anterior and posterior lobe and pituitary stalk in the form of hemorrhage, necrosis, and fibrosis.

Harper et al³⁵ observed pituitary infarcts in 38 of 100 consecutive patients who died as a result of nonmissile head injuries. In this investigation, all patients with large- or medium-sized pituitary infarctions had increased intracranial pressure at some point. Interruption of the hypothalamohypophysial portal blood supply (ie, due to increased intracranial pressure) was assumed to be 1 possible mechanism of anterior lobe infarction.^31,35 Another study⁵⁵ confirmed acute infarction as the underlying adrenohypophysial pathology in those patients with TBI who did not die instantly after the trauma. In 13 of the 30 patients with pituitary specimens who survived TBI between 3 hours and 7 days, acute infarcts of varying size (up to subtotal necrosis of the anterior lobe) were noticed, whereas no infarction was observed in pituitary glands of patients who died immediately (n = 12). Systematic neuropathological investigations of the pituitary gland after aneurysmal SAH have not been conducted but hemorrhages in the pituitary were described in 1 case report.⁵⁶

In another case series, hypothalamic lesions were observed³⁴ in 45 of 106 cases (42%) consisting of hemorrhage (68.9%), necrosis (57.8%), or a combination of both (n = 12; 26.7%). An association of hypothalamic lesions after TBI with temporoparietal blows and fractures of the middle cranial fossa was observed in this study. This connection between fracture of the middle cranial fossa and anterior pituitary necrosis was noted in another study.⁵⁷

In the only systematic neuropathological study specifically on hypothalamic lesions in the wake of SAH, ischemic necrosis, and macrohemorrhages and microhemorrhages in the hypothalamus were seen in 68% of 102 patients.³⁶ In the case of aneurysms close to the midline, bilateral hypothalamic lesions were often observed. The hypothalamic microhemorrhages were described in this study³⁶ as “remarkably selective in their site and surprisingly localized to the paraventricular and supra-optic nuclei, often rendering these nuclei prominent to the naked eye.” As a possible explanation for these selective hemorrhages, a temporary obstruction of venous drainage in these particularly densely vascularized nuclei due to increased pressure in the chiasmatic cistern after the hemorrhage was postulated. Further possible mechanisms of hypothalamic damage after aneurysmal SAH include direct damage of the fine perforating hypothalamic arteries in the subarachnoid space by the hemorrhage, vasoconstriction resulting in ischemic damage, and subarachnoid blood forced up the sheaths of the perforating arteries and then rupturing out into the cerebral parenchyma.³⁶

In summary, large neuropathological series demonstrate pituitary as well as hypothalamic lesions after TBI. In 2 studies, an association of fractures of the middle cranial fossa and such lesions was noted. The scarcer neuropathological evidence on the hypothalamopituitary system after SAH highlights the large degree of hypothalamic damage in patients with SAH.

As mentioned previously, there is a wide variation of the frequencies of hormone deficits reported. The reliability of the methodological tools used to assess pituitary function may be an important factor in this regard. In general, assessment of the growth hormone and adrenocorticotropic hormone axes require dynamic stimulation tests to distinctly separate normal from deficient responses and appropriate cut-offs should be defined considering potential confounding influences of assays used, laboratory tests, age, body mass index, and sex (reviewed in the study by Schneider et al³⁰). Therefore, differences in the reported frequencies may be due to more stringent diagnostic criteria applied by some researchers but not others.

The insulin tolerance test evaluates the integrity of both hypothalamic and pituitary function as opposed to many other tests and has been considered the criterion standard for the evaluation of assessing the growth hormone and the adrenal axes. However, it cannot be performed in patients with severe cardiovascular disease and uncontrolled epileptic seizures, limiting its use in patients with TBI and SAH. The insulin tolerance test has been used by some authors^{10,15,26- 27} with no adverse effects, but other authors have used alternative tests (Table 2). Different cut-off levels used for these tests and different hormone assays might have had an important influence on the frequency of patients defined as hormone deficient.⁵⁸ In addition, some authors used 2 dynamic tests to confirm abnormalities in pituitary function¹⁵ while in other cases only 1 test was used.⁵⁹ Therefore, the robustness of the methods used to diagnose hypopituitarism vary between studies.

In addition, publication bias and patient selection bias—with studies reporting higher frequencies of hypopituitarism being more likely to be published—cannot be completely ruled out. However, this seems unlikely because those studies reporting the lowest frequencies have been published in highly cited journals. Moreover, the fact that hypopituitarism has been found more often in brain-injured patients than in control individuals and the fact that patients with posttraumatic hypopituitarism show the same impairments as patients with other forms of hypopituitarism underline the importance of the problem.

Patients with hypopituitarism require replacement of the deficient hormone as part of their standard clinical care. Adequate hormone replacement can, in general, reverse the symptoms of hypopituitarism and normalize the risks associated with it.³⁰ In patients with brain injury, however, damage of the pituitary may be subtle and sometimes only borderline endocrine disturbances are present. In addition, these patients often have multiple other sequelae of the trauma such as depression, neuropsychological deficits, or personality changes due to organic psychosyndrome. Thus, it is not clear if these patients benefit from hormone replacement in the same way as patients with classic causes of hypopituitarism. Due to the potential serious consequences of corticotrophic, thyrotrophic, or posterior pituitary dysfunction, it is important to adequately treat these patients if convincing biochemical and clinical evidence of these deficiencies is present. Gonadotrophic hormone deficiency is often transient in the early period after brain injury.^{14- 15,21- 22} To date, there is no clear evidence for replacement of sex steroids in the acute phase of TBI or SAH. Although, if hypogonadism persists into the chronic phase, sex hormone replacement should be considered as in all other etiologies of hypopituitarism.

However, prospective, randomized studies are needed to assess the effects of replacement of these hormones in the case of subtle endocrine abnormalities, transient endocrine changes, and when clinical features of deficiency are unclear.

Growth hormone replacement is indicated for adults with severe growth hormone deficiency⁶⁰ and its salutary effects are well documented.⁶¹ In the presence of biochemical and clinical evidence of persistent severe growth hormone deficiency and additional pituitary hormone deficiencies, growth hormone substitution should be considered in patients with TBI and SAH. Yet further studies assessing the benefit of growth hormone replacement, particularly on recovery, rehabilitation, body composition, and neuropsychiatric function in this group of patients are needed.

Considering the large number of individuals who have TBI and SAH each year, posttraumatic hypopituitarism is of major public health importance. Based on the incidence of patients hospitalized for TBI and SAH reported in the literature and the frequencies of hypopituitarism in these patients, we have previously estimated the incidence of hypopituitarism caused by these disorders to be more than 30 patients per 100 000 population per year.³⁰ This by far outnumbers all other well-recognized causes of hypopituitarism.⁶² Due to the previously mentioned limitations of the studies and a potential selection bias, this is only a rough estimate and should be considered with caution. However, it is clear that a large number of patients with hypopituitarism after TBI or SAH remain undiagnosed and untreated.

Therefore, integrated screening programs for posttraumatic hypopituitarism should be developed and incorporated as standard clinical care for the patient with acute brain injury from trauma or SAH. Close collaboration between neurosurgery, endocrinology, rehabilitation medicine, and other interested disciplines is essential to ensure optimal delivery of care

Early posttraumatic pituitary dysfunction can be transient in many cases and conversely, hypopituitarism can evolve over several weeks or months after injury.^{14,19,21- 22} Therefore, periodic evaluation in the first year after trauma may be necessary.

In the acute phase of brain injury, the diagnosis of adrenal insufficiency should not be missed because it can be life threatening.^42- 43 Patients should be screened for signs and symptoms of hypocortisolism including hyponatremia, hypotension, and hypoglycemia. Because dynamic assessment of adrenocorticotropic hormone reserve is not practical under conditions of acute critical illness, we suggest that morning serum cortisol concentrations be checked in the first days after trauma or SAH.⁴³ Defining a cortisol cut-off that will help diagnose adrenal failure in acute illness is difficult because serum total cortisol values under such conditions will be influenced by several factors including the degree of severity of the underlying illness, sepsis, and medications. Serum levels of cortisol-binding globulin can be reduced in catabolic states resulting in disproportionately low total cortisol compared with free (biologically active) cortisol (clinical laboratory tests that only measure total cortisol as free cortisol is technically difficult, time-consuming, and expensive). Allowing for these confounding factors, acute-phase morning cortisol level of less than 7.2 μg/dL (200 nmol/L) may be suggestive of adrenal insufficiency in acutely ill patients with TBI or SAH, and glucocorticoid replacement should be instituted.²⁰ However, values between 7.2 and 18 μg/dL (200-500 nmol/L) in the presence of features suggestive of adrenal insufficiency such as hyponatremia, hypoglycemia, hypotension, or unexpected slow recovery may still be inappropriately low and a trial of glucocorticoid therapy should be considered.

Assessment of the growth hormone, gonadal, and thyroid axes is not necessary in the acute phase because there is currently no evidence that acute-phase therapy with these hormones improves outcome. Between 3 and 6 months after injury, all patients should undergo careful screening for clinical signs of hypopituitarism. Particular attention should be paid to loss of secondary hair, new oligomenorrhea or amenorrhea, impaired sexual function, weight changes, polydipsia, the above-mentioned signs of hypocorticalism, and poor recovery. If any of these signs is present, pituitary assessment should be performed. Because the sequelae of brain injury may mask the signs of hypopituitarism, the threshold for endocrine assessment should be low and in cases of uncertainty, endocrine assessment should be performed at least once. Also, in patients with basal skull fractures, diffuse axonal injury, increased intracranial pressure, or prolonged intensive care unit stay, pituitary assessment should be considered. If hypopituitarism is detected, hormone therapy should be given as appropriate. In patients with documented anterior hypopituitarism at 3 to 6 months postinjury, repeat anterior pituitary assessment at 1 year may be considered if the clinical or biochemical parameters raise the possibility of delayed recovery.

In conclusion, hypopituitarism is a common, potentially serious but treatable complication of TBI and SAH. Increased level of awareness among physicians of all disciplines who are involved in the care of patients with TBI and SAH is required to identify affected cases and provide the appropriate and timely hormone therapy, which has the potential to improve recovery, rehabilitation, and quality of life for those patients.