The optimal hyperosmotic agent for treatment of intracranial hypertension following TBI has been an inexhaustible subject of debate. Scant evidence underlies these discussions despite widespread and long-standing use of both agents. In this article Anstey et al. retrospectively analysed data from 14 different centres to compare outcomes in mannitol-only and hypertonic saline-only cohorts. They excluded those who had received both agents and focused on early administration (the first 96 hours), which yielded 46 patients in each group. The results were quite surprising in that they showed such a notable difference in mortality (primary outcome), with a hazard ratio the mannitol-only group of 2,64. This was after adjustment for key mortality predictors. The study is small, hampered by a few between-group differences in baseline characteristics and naturally limited by its retrospective design. Nonetheless, the large difference in effect-size is cause for some concern about the role of mannitol and underscores the need for a robust prospective investigation to further inform the choice of agent.