Prevalence and predictive factors of post-traumatic hypopituitarism

Objective To estimate the prevalence and predictive factors of hypopituitarism following traumatic brain injury (TBI). Design A cross-sectional cohort study. Patients One hundred and four hospitalized TBI patients (26F/78M), median age 41 (range 18-64) years, body mass index (BMI) 25 (17-39) kg/m(2); severity: mild [Glasgow Coma Scale (GCS) score 13-15) n = 44, moderate (GCS 9-12) n = 20, severe (GCS < 9) n = 40]. Measurements Patients were evaluated 13 (10-27) months post-injury, with measurement of baseline (0800-1000 h) and post-stimulatory hormonal levels during an insulin tolerance test (ITT) (86%) or, if contraindicated, an arginine(arg)-GHRH test + Synacthen test (14%). Insufficiencies were confirmed by retesting. Results Hypopituitarism was found in 16 (15%) patients, affecting one axis in 10, two axes in four and more than two axes in two patients. The GH axis was most frequently affected (15%), followed by secondary hypoadrenalism (5%), hypogonadism (2%), hypothyroidism (2%) and diabetes insipidus (2%). The risk of pituitary insufficiency was increased in patients with severe TBI as opposed to mild TBI [odds ratio (OR) 10.1, 95% confidence interval (CI) 2.1-48.4, P = 0.004], and in those patients with increased intracerebral pressure [OR 6.5, 95% CI 1.0-42.2, P = 0.03]. Patients with only one affected axis were all GH deficient; 60% (n = 6) of these were overweight or obese. Conclusions The prevalence of hypopituitarism was estimated at 16%. Although high, this value was lower than previously reported, and may still be overestimated because of well-known confounding factors, such as obesity. Indicators of increased TBI severity were predictive of hypopituitarism, with a high negative predictive value. Neuroendocrine evaluation should therefore be considered in patients with severe TBI, and in particular in those with increased intracerebral pressure (ICP).