Cortisol evaluation during the acute phase of traumatic brain injury - a prospective study.

Background: Biochemical diagnosis of adrenal insufficiency (AI) is difficult in the context of traumatic brain injury (TBI). Aim: To assess the frequency and predictive factors of AI in victims of TBI from Algiers. Methods: Between November 2009 and December 2013, TBI victims had a single 8-9 am serum cortisol measurement during the acute postinjury period (0-7 days). AI was defined according to basal cortisol levels of 83, 276 and 414 nmol/L. Variables studied were TBI severity according to Glasgow coma scale, duration of intubation and coma, pupillary status, hypotension, anaemia, brain imaging findings, diabetes insipidus and medication. Insulin tolerance test was performed during the recovery phase, defining AI as peak cortisol <500 nmol/L. Results: Cortisol samples were obtained at median 3 (1-7) days from 277 patients (257M: 20F) aged 32 (18-65) years. Acute AI frequency was 8 (2.8%), 20 (21%) and 35 (37%), respectively using the three cortisol cut-offs. Factors predicting AI were diastolic hypotension, sedative medication, diabetes insipidus, skull base fracture and intraparenchymal haematoma. Mortality was highest in patients with acute cortisol <276 nmol/L (44.6% with OR for death 1.64, 95% CI 0.92-3.0, P = .12). During the recovery phase, AI was present in 3 of 3, 12 of 24, 4 of 16 and 20 of 66 patients with week 1 cortisol <83, 83-276, 277-414 and >414 nmol/L. Conclusion: Hydrocortisone replacement is advised in TBI patients with morning cortisol <276 nmol/L or those < 414 nmol/L with additional risk factors for AI. As acute and subsequent AI are poorly correlated, patients with moderate/severe TBI require adrenal re-evaluation during the recovery phase.