In order to identify whether low-dose (1 microg) tetracosactide (Synacthen) testing may be preferable to high-dose (250 microg) testing in the diagnosis of adrenal insufficiency in traumatic brain injury (TBI), as suggested by studies in other forms of critical illness. We retrospectively reviewed the results of modified tetracosactide tests (involving administration of both low-dose and high-dose tetracosactide) conducted for clinical indications in patients in a neurocritical care unit within 10 days of TBI. Sixty-three modified tests were included and cortisol concentrations before and after administration of tetracosactide were extracted from the hospital records. Data were also extracted regarding hemodynamic response to empirical corticosteroid therapy, based on rapid weaning from vasoactive drugs. Cortisol increments at 30 and 60 min following tetracosactide correlated well in the low-dose test (r(2) = 0.875, P < 0.0001). The mean cortisol concentration was 581 nmol/l at 30 min and 556 nmol/l at 60 min in the low-dose test. Cortisol increments following low-dose and high-dose testing correlated well overall (r(2) = 0.839, P < 0.0001), but results were discordant in 27 of 63 cases (43%) when the same diagnostic threshold was used. ROC curve analysis showed that both tests performed poorly in identifying hemodynamic steroid responsiveness (AUC 0.553 and 0.502, respectively). In the low-dose tetracosactide test, it is sufficient to determine cortisol concentrations at baseline and at 30 min. Low-dose and high-dose tests give discordant results in a significant proportion of cases when using the same diagnostic threshold. Neither test can be used to guide the initiation of corticosteroid therapy in acute TBI.