Study	Reason for exclusion
Battison 2005	This is a prospective, cross‐over, pilot trial.
Bourdeaux 2011	Participants received doses of HTS or sodium bicarbonate in a random order, some having both. This may result in cross‐over effects between treatments.
Du 2017	This trial was excluded because there was no defined trigger for starting hyperosmotherapy to reduce ICP, and mannitol was administered every eight hours.
Hong 2017	This trial was excluded due to unreliable methods (HTS or other ICP‐lowering agents were administered by time but not according to ICP monitoring).
Huang 2014	This is a randomised clinical trial with a cross‐over design.
Huang 2015	This trial was excluded for questionable sequence generation; there was no reporting on outcomes of interest to this review.
Ichai 2009	Sodium lactate was used to decrease the raised ICP. Sodium lactate differs fundamentally from sodium chloride; infusion of sodium lactate in TBI decreases the occurrence of raised ICP. The absence of significant modification of plasma osmolarity does not support a pure osmotic effect of sodium lactate.
Jafari 2018	This trial was excluded because it assessed the effects of adding furosemide to 1 of 2 groups both receiving HTS.
Jiang 2018	RCT. Excluded because of ineligible comparator (there were 3 arms, all of HTS, in the same concentration but delivered at different speeds (6 mL/3 mL/2 mL per h).
Jin 2018	Excluded due to doubts about sequence generation and apparently unreliable data. We were unable to contact trial authors for clarification.
Li 2018	Excluded due to unreliable data and methods (there is an error in the judgement of P value).
Liang 2013	Excluded due to unreliable data and methods (incorrect 'T' values, questionable sequence generation). We were unable to contact trial authors for clarification.
Liu 2018	Excluded due to unreliable methods (HTS or other ICP‐lowering agents were administered by time but not according to ICP monitoring).
Mei 2016	Excluded due to unreliable data and methods (incorrect 'T' values, questionable sequence generation). We were unable to contact trial authors for clarification.
NCT01028339	Appeared to meet inclusion criteria but according to the clinicaltrials.gov record it has been "terminated. (No patients enrolled during 2 years)".
NCT01108744	This trial met inclusion criteria but according to the clinicaltrials.gov record was "...withdrawn prior to enrolment. Timeline to consent prior to intervention start was unfeasible".
NCT01111682	This trial met inclusion criteria but according to the clinicaltrials.gov record was "...terminated. (A significant reduction in head injuries coupled with more frequent use of craniectomy reduced the number of potential subjects)".
NCT01215019	This trial met inclusion criteria but according to the clinicaltrials.gov record was "...withdrawn prior to enrolment. (Lack of funding; no subjects enrolled)".
Ni 2018	Excluded due to unreliable data and methods (e.g. no clear definition of ICP).
Polushin 2009	This is a randomised clinical trial with a cross‐over design.
Roquilly 2017	This planned RCT is measuring continuous hyperosomolar therapy (HTS) in participants with TBI vs standard care (which is unlikely to include other ICP‐lowering agents).
Sakellaridis 2011	This is a prospective, cross‐over, pilot trial.
Shu 2015	Excluded due to unreliable data and methods (incorrect 'T' values, questionable sequence generation). We were unable to contact trial authors for clarification.
Upadhyay 2010	The sequence generation was described (quasi‐randomisation). The included participants were 200 children with increased ICP of various different aetiologies, of whom only 14 had TBI. Comparison of average reduction of MAP at defined intervals was performed to indirectly assess reduction in ICP.
Wang 2017	This trial was non‐therapeutic in nature, measuring only coagulation.
Yang 2019	This trial was excluded because it compared two different doses of HTS (7.5% HTS (4 ml/kg) and 3% HS treatment (4 ml/kg)).
Zhang 2014	Excluded due to unreliable data and methods (incorrect 'T' values, questionable sequence generation). We were unable to contact trial authors for clarification.
Zhang 2015	Excluded due to unreliable data and methods (incorrect 'T' values, questionable sequence generation). We were unable to contact trial authors for clarification.
Zhang 2018	Excluded due to unreliable methods (HTS or other ICP‐lowering agents were administered by time, but not according to results of ICP monitoring).

HTS: hypertonic saline; ICP: intracranial pressure; MAP: mean arterial pressure; RCT: randomised controlled trial; TBI: traumatic brain injury;