ICP managment

It found no evidence of an effect on any clinically important outcome. Glasgow Outcome Scale at six months showed no difference (risk ratio 0.82, 95% CI 0.48 to 1.40), nor did all-cause mortality (risk ratio 0.96, 95% CI 0.60 to 1.55) or total length of stay. The point estimate for uncontrolled intracranial pressure favoured hypertonic saline but did not reach significance (risk ratio 0.52, 95% CI 0.26 to 1.04). The harm identified was hypernatraemia, roughly twice as common with hypertonic saline (risk ratio 2.13, 95% CI 1.09 to 4.17). Most of the included trials were at unclear or high risk of bias and the evidence was rated low to very low certainty.

Set this beside the earlier meta-analyses and the shape of the evidence becomes clear: hypertonic saline moves intracranial pressure and cerebral perfusion pressure in some trials and at some time points, has never been shown to change six-month outcome, and carries a measurable sodium cost. This review is the most conservative of the three — here even the reduction in uncontrolled intracranial pressure did not reach statistical significance, on low to very low certainty evidence. That is enough to justify keeping it in the algorithm and not enough to justify claiming it is better.

A. 1.16 (95% CI 1.00 to 1.33) favouring hypertonic saline, with a mean intracranial pressure difference of 2.0 mmHg (95% CI -1.6 to 5.7)

Note how narrow the margin is. The lower bound of the relative risk sits on 1.00 and the mean intracranial pressure difference crosses zero. The authors' own reading is that hypertonic saline may be superior and that a large multicentre trial is needed to settle first-line therapy — a 16% relative advantage in five small trials is a reason to keep hypertonic saline in the algorithm, not a reason to say mannitol has been beaten.

The physiological endpoints favoured hypertonic saline, but only after a delay: intracranial pressure was no different at 30 to 60 minutes (mean difference -0.19 mmHg) and significantly lower after hypertonic saline at 90 to 120 minutes (mean difference -2.33 mmHg, 95% CI -3.17 to -1.50). Cerebral perfusion pressure was higher after hypertonic saline at both windows, by 5.48 and 9.08 mmHg. The patient-centred outcomes showed nothing: mortality did not differ (relative risk 0.69, 95% CI 0.45 to 1.04) and neither did favourable neurological outcome (relative risk 1.28, 95% CI 0.86 to 1.90). The trial sequential analysis added the crucial qualifier — the accumulated number of patients is insufficient to make reliable statements about long-term outcome, so the absence of an outcome difference here is an underpowered result rather than a demonstrated equivalence.

A drug that lowers the number on the monitor has not yet been shown to change what happens to the patient. Trial sequential analysis is the tool that tells you whether "no difference" has actually been tested or merely not yet observed.