Discussion
Summary of Evidence
In this meta-analysis of six randomised trials, no statistically or clinically significant difference was found between hypertonic sodium solutions and mannitol in lowering ICP in TBI. There was a trend towards hypertonic sodium solutions lowering ICP more effectively but the pooled weighted CI included the value of no difference. Individually, one trial showed a statistically greater quantitative ICP reduction using hypertonic sodium than with mannitol, one showed a trend in favour of hypertonic sodium, three studies showed equivalence and one study suggested a trend in favour of mannitol. Even the largest study was not deemed to have sufficient power for definite conclusions to be reached. This meta-analysis supports the trend in clinical practice, that there is no statistically significant difference in the ability of mannitol and hypertonic sodium solutions to effectively lower ICP in TBI.
Limitations
The following methodological differences between studies existed:
formulation of hypertonic sodium solutions;
dosage used;
varying thresholds of raised ICP at which treatment was initiated;
inclusion of non-TBI neurosurgical patients (in two studies);
the time at which outcomes were assessed.
This heterogeneity is reflected in the I value of 38.1%, and is the basis upon which a random-effects model was used. Only one study suggested that mannitol is more favourable than hypertonic sodium solutions. However, this trial applied a host of exclusion criteria both during recruitment and intervention, introducing significant selection bias and precluding an intention-to-treat analysis. The best designed study, by Ichai et al, was a manufacturer-sponsored study and was published without a conflict of interest statement. This is also the case for the study by Battison et al.
What This Study Adds
Another recent meta-analysis of hypertonic saline versus mannitol for the treatment of raised ICP, which included all patients with raised ICP (including, but not exclusively, patients with TBI) from multiple causes, found that hypertonic sodium solutions were more effective than mannitol at controlling episodes of raised ICP. As the predominant mechanism of cerebral oedema (cytotoxic versus vasogenic) may vary between aetiologies and hence the treatment mechanism may be different, we considered it important to limit our cohort to patients with TBI as far as possible. Mortazavi et al reviewed the literature on hypertonic saline for treating raised ICP of all causes. While their meta-analysis involved a less rigorous search, they essentially agreed that the majority of studies showed a more favourable short-term ICP outcome for hypertonic saline. For patient focus and neurological outcome, Mortazavi found no clear benefit of hypertonic saline over mannitol but declared a minor positive trend for hypertonic saline.
Three observational studies support the finding that both hypertonic saline and mannitol work to effectively reduce ICP in TBI, with a suggestion that hypertonic saline might be more effective in this respect. Two randomised trials, which did not meet our inclusion criteria, also support the use of hypertonic sodium solutions in patients with TBI with raised ICP. Two other meta-analyses have been published in this field but one was methodologically flawed and the other reviewed the role of the two agents in patients with raised ICP from any cause. This meta-analysis provides a methodologically secure endorsement of the use of hypertonic normal saline in the reduction of ICP in patients with TBI.
Future Research Agenda
Future research into the use of hyperosmolar agents, given as part of a care bundle in the emergency department, is necessary. A multicentre randomised controlled trial with sufficient power to detect any clinically meaningful difference in neurological recovery, and surrogate outcomes such as ICP measurements, would be required. Until we have such data, the body of evidence obtained in neuro-intensive care units should be used to guide best practice in the emergency department.