Hypopituitarism Following Traumatic Brain Injury: Determining Factors for Diagnosis

Neuroendocrine dysfunction, long recognized as a consequence of traumatic brain injury (TBI), is a major cause of disability that includes physical and psychological involvement with long-term cognitive, behavioral, and social changes. There is no standard procedure regarding at what time after trauma the diagnosis should be made. Also there is uncertainty on defining the best methods for diagnosis and testing and what types of patients should be selected for screening. Common criteria for evaluating these patients are required on account of the high prevalence of TBI worldwide and the potential new cases of hypopituitarism. The aim of this review is to clarify, based on the evidence, when endocrine assessment should be performed after TBI and which patients should be evaluated. Additional studies are still needed to know the impact of post-traumatic hypopituitarism and to assess the impact of hormone replacement in the prognosis.

Neuroendocrine dysfunction, long recognized as a consequence of traumatic brain injury (TBI), is a major cause of disability that includes physical and psychological involvement with long-term cognitive, behavioral, and social changes. There is no standard procedure regarding at what time after trauma the diagnosis should be made. Also there is uncertainty on defining the best methods for diagnosis and testing and what types of patients should be selected for screening. Common criteria for evaluating these patients are required on account of the high prevalence of TBI worldwide and the potential new cases of hypopituitarism. The aim of this review is to clarify, based on the evidence, when endocrine assessment should be performed after TBI and which patients should be evaluated. Additional studies are still needed to know the impact of post-traumatic hypopituitarism and to assess the impact of hormone replacement in the prognosis.
Keywords: hypopituitarism, traumatic brain injury, growth hormone deficiency, hypogonadism Traumatic brain injury (TBI) is a significant public health problem worldwide, with an overall incidence in Western countries of 200-235 cases per 100,000 individuals annually (Tagliaferri et al., 2006). Approximately 20-40% of TBI episodes involve moderate to severe traumatic injuries (Frankowski et al., 1985), while 10% of cases are fatal.
In young adults TBI is considered one of the main causes of death (Hodgson et al., 2000). Moreover, it is a major cause of disability that includes physical and psychological involvement with long-term cognitive, behavioral, and social changes (Salazar et al., 2000;Khan et al., 2003). Neuroendocrine dysfunction has long been recognized as a consequence of TBI (Cyran, 1914), particularly in moderate and severe cases, although the last 10 years have witnessed a better understanding of the frequency of TBImediated hypopituitarism and its clinical implications, with the publication of an increasing number of studies (Benvega et al., 2000;Kelly et al., 2000;Lieberman et al., 2001;Agha et al., 2004;Bondanelli et al., 2004;Popovic et al., 2004;Aimaretti et al., 2005;Leal-Cerro et al., 2005;Herrmann et al., 2006;Schneider et al., 2006;Tanriverdi et al., 2006). This review is focused in adult patients, but there is also recent data about TBI-induced hypopituitarism in children and adolescents (Khadr et al., 2010;Moon et al., 2010;Norwood et al., 2010).

EPIDEMIOLOGY OF HYPOPITUITARISM
Despite the high prevalence of TBI in the general population, brain injury-related hypopituitarism remains largely under-diagnosed Gasco et al., 2010). Brain damage often presents with signs and symptoms that can easily mask the typical symptoms of hypopituitarism (Hellawell and Pentland, 2001), and physicians who care for such patients have little awareness about the possibility of pituitary deficiencies and their impact on the prognosis, resulting in misdiagnosed cases as an endocrine assessment is not usually considered.
To date, the prevalence of reported hypopituitarism following TBI varies widely among published studies (Kelly et al., 2000;Lieberman et al., 2001;Popovic et al., 2004;Leal-Cerro et al., 2005;Tanriverdi et al., 2006). Considering all degrees of hypopituitarism, from isolated pituitary deficiencies to multiple or even total hypopituitarism, the prevalence ranges from 15 to 90% (Kelly et al., 2000;Lieberman et al., 2001;Popovic et al., 2004;Leal-Cerro et al., 2005;Tanriverdi et al., 2006). A meta-analysis including 1015 patients with TBI reported a pooled prevalence of hypopituitarism of 27.5%. The prevalence of anterior pituitary dysfunction in the studies included in this meta-analysis ranged from 15 to 68%. The design and time of pituitary evaluation in these studies were also variable, ranging from 3 months to 7 years following the head trauma (Schneider et al., 2007).

PATHOPHYSIOLOGY OF HYPOPITUITARISM
The pathophysiology of hypopituitarism following TBI is not completely understood, and several factors implicated in its development have been suggested (Dusick et al., 2008).
An important mechanism involved is infarction or ischemia of the pituitary gland (Malya et al., 2008), an explanation also supported by the high prevalence of hypopituitarism in patients with www.frontiersin.org ischemic stroke (Bondanelli et al., 2010). Vascular damage to the pituitary can also be caused by several mechanisms: (1) Traumatic damage to the long hypophyseal portal veins and subsequent venous infarction are thought to be the main underlying vascular mechanisms (Daniel et al., 1959;Kornblum and Fisher, 1969).
(2) Direct traumatic injury to the pituitary gland as demonstrated by the high prevalence of necrosis or hemorrhage into the pituitary gland shown in postmortem studies (Crompton, 1971;Salehi et al., 2007). (3) Multiple secondary insults from hypotension, hypoxia, anemia, and brain swelling, which could also lead to an ischemic pituitary gland (Wachter et al., 2009). (4) Another mechanism described as responsible for hypopituitarism due to vascular damage is transection of the pituitary stalk, which potentially causes hypopituitarism because of infarction of the pituitary tissue (Daniel et al., 1959;Malya et al., 2008).
However, recent research has indicated a possible interaction between autoimmunity and the development of hypopituitarism following TBI. It has been demonstrated that antipituitary antibodies are present in patients with TBI-induced pituitary dysfunction and persist even 3 years after diagnosis (Tanriverdi et al., 2008a(Tanriverdi et al., , 2010a.

WHICH ENDOCRINE DEFICITS ARE THE MOST FREQUENT?
The degree of hypopituitarism, considered as the number of axes involved, and the prevalence of each pituitary deficiency differ in the various studies published. Some authors have identified central hypogonadism as the most frequent pituitary deficiency after TBI (Benvega et al., 2000;Tanriverdi et al., 2006;Wachter et al., 2009), although most have reported the GH deficiency (GHD) as the most prevalent (Bondanelli et al., 2004;Popovic et al., 2004;Aimaretti et al., 2005). Taking into consideration all data published to date, GHD is the most common hormone deficiency in both severe and moderate TBI. However, its prevalence varied from 2 to 66%, with 39% of cases of severe GHD. Prevalence of the other pituitary deficiencies was also variable among the studies and ranged from 0 to 60% for secondary adrenal insufficiency, 0 to 29% for secondary hypothyroidism, 0 to 29% for central hypogonadism, and 0 to 48% for abnormal hyperprolactinemia (Benvega et al., 2000;Kelly et al., 2000;Lieberman et al., 2001;Agha et al., 2004;Bondanelli et al., 2004;Popovic et al., 2004;Aimaretti et al., 2005;Leal-Cerro et al., 2005;Herrmann et al., 2006;Schneider et al., 2006;Tanriverdi et al., 2006;Kokshoorn et al., 2010). All cases of abnormal hyperprolactinemia were mild elevations of prolactin and had no clinical significance (Kokshoorn et al., 2010).
This variability could be explained by several factors, including a different time interval between TBI and the assessment of pituitary function (from 24 h to 23 years), the type and severity of brain injury, and the different methods used to evaluate pituitary function reserve, such as endocrine tests, assays, and criteria for the diagnosis of hypopituitarism (Gasco et al., 2010;Kokshoorn et al., 2010).
The study design might also influence the prevalence of hypopituitarism. Only a few studies have been designed to evaluate the incidence of hypopituitarism following TBI with specific criteria and evaluation at a particular time (Aimaretti et al., 2004Tanriverdi et al., 2006Tanriverdi et al., , 2008b and most of the studies published are retrospective or cross-sectional (Kelly et al., 2000;Lieberman et al., 2001;Agha et al., 2004;Bondanelli et al., 2004;Popovic et al., 2004;Leal-Cerro et al., 2005;Herrmann et al., 2006).

WHEN SHOULD PATIENTS BE SCREENED AFTER TBI?
The time interval between TBI and pituitary hormone assessment is one of the major factors responsible for variations in the prevalence of hypopituitarism.
The presence of pituitary dysfunction in the early acute phases (first 24 h) or acute phases (up to 2-3 weeks) following TBI was not related to hypopituitarism after 12 months (Tanriverdi et al., 2006(Tanriverdi et al., , 2008b, as pituitary function was recovered in many patients (Tanriverdi et al., 2008b). Indeed, pituitary deficiencies were found in 56% of patients when endocrine evaluation was performed 3 months after TBI. However, when it was performed 12 months after TBI, pituitary deficiencies were only detected in 36% of cases. GHD was detected in 9 and 10%, at 3 and 12 months after TBI, respectively. Hypogonadism was detected in 32 and 20%, ACTH deficiency in 19 and 9%, and TSH deficiency in 8 and 3%, respectively (Schneider et al., 2006).
Prospective studies have shown the dynamic property of hypopituitarism in isolated and multiple pituitary deficiencies as well as in the worsening or recovery of pituitary function Tanriverdi et al., 2006Tanriverdi et al., , 2008b. Pituitary recovery 12 years after TBI was also reported (Ruggeri et al., 2010), possibly due to pituitary revascularization and cellular pituicyte repopulation (Benvega et al., 2000).
In the early post-traumatic period there are often hormonal changes that mimic pituitary dysfunction (Klose et al., 2007a). This is included in the physiological response to acute and critical illness, which comprises hormonal changes similar to GHD, central hypogonadism, and hypothyroidism (Van den et al., 1998). Acute illness or drugs frequently used in these clinical conditions could alter the metabolism of the protein binding hormones, resulting in false deficiencies (Agha et al., 2005a).
On the other hand, stress hormones such as cortisol, prolactin, and vasopressin are increased (Hamrahian et al., 2004;Jochberger et al., 2006). Consequently, as cortisol should be increased in the acute phase after TBI, low or low-normal baseline free cortisol could still be proposed as a potential early marker for the development of chronic ACTH deficiency in head trauma patients (Klose et al., 2007a).
This overestimation of hypopituitarism in the acute phase following TBI makes it reasonable to propose the evaluation of pituitary function at least 1 year after the head trauma and not at an earlier time point, particularly for hormones that do not necessarily need to be replaced urgently, such as the gonadal and somatotropic axes (Lorenzo et al., 2005).

WHICH PATIENTS SHOULD BE SELECTED FOR SCREENING?
Severity of head trauma has been proposed as a predictor of occurrence of hypopituitarism and is evaluated by the Glasgow Coma Scale (GCS) Score (Teasdale and Jennett, 1974) according to the consciousness level and ocular and motor movements.
We propose an algorithm (Figure 1) to know who are the patients that should be screened for pituitary function, based on evidence.
Many studies (Kelly et al., 2000;Lieberman et al., 2001;Schneider et al., 2006;Tanriverdi et al., 2006) showed that pituitary deficiencies were more frequent in cases of moderate and severe head trauma (GCS < 13), although patients with severe disability, as vegetative states, will not benefit from hormonal replacement and should be excluded from non-vital deficiency assessment. On the other hand, other authors did not find any association between injury severity and prevalence of hypopituitarism (Agha et al., 2004;Aimaretti et al., 2005;Leal-Cerro et al., 2005;Tanriverdi et al., 2006). In addition, low-intensity repeated trauma, which occurs in certain sports, may also result in pituitary deficiencies (Tanriverdi et al., 2010b) probably related to a decrease in pituitary volume (Tanriverdi et al., 2008c).
Mild TBI, defined as a score between 13 and 15 in the GCS, is associated with more frequent isolated pituitary deficiencies with a higher proportion of axes recovered after 1 year of follow-up (Tanriverdi et al., 2006). Conversely, studies published to date collectively indicate that at 6-2 months post-injury, 25-40% of moderate, and severe TBI patients will develop chronic hypopituitarism with the somatotropic and gonadotropic axes being most commonly affected and with a high proportion of persistent deficiencies (Schneider et al., 2006;Tanriverdi et al., 2006Tanriverdi et al., , 2008b. Indeed, the pooled prevalence of hypopituitarism in cases of severe, moderate and mild TBI reported by Schneider et al. (2007) was 35.3, 10.9, and 16.8%, respectively.
There is no cost/benefit indication for global screening of all cases of mild TBI . However, some situations, such as the presence of anatomic abnormalities on initial CT (diffuse axonal injury or skull fractures; Schneider et al., 2008), diffuse brain swelling, hypoxia, hypotension (Kelly et al., 2000), duration of coma (Benvega et al., 2000), increased intracranial pressure, longer intubation and hospitalization, and advanced age have been associated with a worse prognosis and higher incidence of hypopituitarism (Becker et al., 1977;Bullock et al., 1996;Hellawell and Pentland, 2001).

WHICH ARE THE MOST APPROPRIATE TESTS FOR ENDOCRINE ASSESSMENT?
Brain injury-related hypopituitarism is under-diagnosed (Schneider et al., 2007;Gasco et al., 2010), but as the health resources of any population are limited, screening every patient who undergoes any head trauma for hypopituitarism should not be recommended. Accordingly, the most important aspect concerning the diagnosis of TBI-associated hypopituitarism is to define the conditions related to an increased risk of hypopituitarism, the time interval from TBI when it is best to evaluate and the endocrine test that should be performed in such cases.
Endocrine tests and criteria for diagnosis vary among the studies published. In the case of the GH-IGF-1 axis all patients were diagnosed with a dynamic test in addition to basal GH and IGF-1 values. However, the dynamic tests used also differ among the studies and include the insulin tolerance test (Kelly et al., 2000;Klose et al., 2007a), the glucagon stimulation test (Agha et al., 2004;Bushnik et al., 2004;Leal-Cerro et al., 2005), the GHRH-arginine test (Bondanelli et al., 2004;Aimaretti et al., 2005;Herrmann et al.,  www.frontiersin.org 2006), and the GHRH-GHRP6 test (Popovic et al., 2004;Tanriverdi et al., 2006). The vagaries of GH provocation test are well known (Ho, 2007). The optimal diagnostic is the one that identifies those patients that will benefit from treatment with GH and excludes those that will not (Blair, 2010). The GHRH-GHRP6 test is rigorous and reliable (Popovic et al., 2000) and is not affected by known factors that modify GH secretion such as obesity. It has demonstrated its utility in the diagnosis of GHD after TBI (Popovic et al., 2004;Tanriverdi et al., 2006) even in a single assay after a 30-min test (Castro et al., 2007), which is essential from a practical point of view. A GH peak of under 10 μg/l will confirm the diagnosis and above 20 μg/l will exclude GHD. In cases between 10 and 20 μg/l another test is needed.
In the case of the adrenal axis, adrenal insufficiency can be diagnosed with a basal level of cortisol when it is extremely low (under 50 nmol/l, 1.8 μg/dl). For the other cases a stimulation test is needed, and the ACTH stimulation (Lieberman et al., 2001;Tanriverdi et al., 2006) and insulin tolerance tests (Kelly et al., 2000;Agha et al., 2005b;Leal-Cerro et al., 2005;Herrmann et al., 2006) are considered the most appropriate for diagnosis with a cut-off cortisol level of 550 nmol/l (20 μg/dl). In the studies previously mentioned, ACTH stimulation test was either performed with 250 μg of cosyntropin (Lieberman et al., 2001) and 1 μg of tetracosactrin (Tanriverdi et al., 2006).
For the thyroid, gonadal, and prolactin axes basal levels are enough for a diagnosis to be made Herrmann et al., 2006;Tanriverdi et al., 2008b). Low levels of free T4 with normal or low TSH allow diagnosis of central hypothyroidism to be made, and low levels of testosterone or estrogens with normal or low FSH and LH are diagnostic of central hypogonadism (Popovic et al., 2004;Herrmann et al., 2006;Klose and Feldt-Rasmussen, 2008). Abnormal prolactin is also diagnosed when basal prolactin levels are not within the normal range (Lieberman et al., 2001;Popovic et al., 2004;Herrmann et al., 2006;Klose and Feldt-Rasmussen, 2008).

OUTCOME
Pituitary function affects TBI patient outcome. Hypopituitarism is thought to contribute to TBI-associated morbidity (Klose and Feldt-Rasmussen, 2008) and to functional and cognitive final outcome (Bondanelli et al., 2007). In addition, GHD has the greatest effect on outcome (Kelly et al., 2006;Bondanelli et al., 2007). Moreover, a clear association has recently emerged between post-TBI neuroendocrine dysfunction and neurobehavioral and quality of life impairment (Klose and Feldt-Rasmussen, 2008), which have increased the interest of doctors in this clinical condition.
Specific evaluations of these patients demonstrated an association between hypopituitarism and especially GHD and depression and memory, neuropsychological and cognitive disorders (Kelly et al., 2006).
Hypopituitarism, when untreated, has also been associated with an adverse lipid profile, increased abdominal fat and body mass index, and deterioration in quality of life (Klose et al., 2007b). It may also be expected that would significantly impede recovery and rehabilitation following TBI (Blair, 2010).
The Pfizer International Metabolic Database (KIMS) reported a significant improvement in quality of life, assessed by the adult growth hormone deficiency assessment (AGHDA) score, in patients with post-traumatic hypopituitarism and GH treatment (Kreitschmann-Andermahr et al., 2008). Recent data also suggests that some of the cognitive impairments observed in patients with GHD after TBI can improve with GH replacement (High et al., 2010). Serious and life-threatening adrenal crises due to ACTH deficiency after TBI have significantly improved with glucocorticoid replacement (Agha et al., 2005b). However, long-term followup studies are still needed to evaluate outcome and the necessity and benefits of hormone replacement with growth hormone and gonadal steroids in these patients.

CONCLUSION
The incidence of hypopituitarism following TBI is around 50 patients per 100,000 individuals per year, which results in a high number of patients affected. Endocrine assessment should be performed 1 year after head injury and only in cases of moderate or severe TBI, or if mild TBI involves conditions related to a worse prognosis. However, additional studies are still needed to assess the impact of hormone replacement in the prognosis.

ACKNOWLEDGMENTS
CIBEROBN is an initiative of the Instituto de Salud Carlos III (ISCIII).