Schwartz 1984.

Methods	Multi‐centre randomised controlled trial.
Participants	People with severe head injury (GCS ≤ 7). Raised ICP (25 torr for more than 15 minutes). 59 people were randomised, stratification was by presence of haematoma. No haematoma (30 people): Pentobarbital 13, Mannitol 17 Haematoma (29 people): Pentobarbital 15, Mannitol 14
Interventions	Pentobarbital versus mannitol. Quote: "Prior to randomisation, people thought to have (but not yet proven to have) raised ICP were given a rapid infusion of 20% mannitol solution." Quote: "All patients were given dexamethasone in an initial dose of 10 mg IV followed by 4 mg q6h." Pentobarbital: Quote: "Loading dose up to 10mg/kg. Maintenance 0.5‐3 mg/kg/hr with serum level monitoring, provided that cerebral perfusion pressure remained above 50 torr. Additional increments of pentobarbital were given to maintain the intracranial pressure at less than 20 torr. The maximum suggested barbiturate level was 45 mg/L. When necessary, dopamine plus volume infusions were administered to raise the systemic arterial blood pressure and hence the cerebral perfusion to at least 50 torr." Mannitol: Quote: "20% solution initial dose 1 g/kg with serum osmolality monitoring. Additional increments of mannitol, usually less than 350cc. were given as required for continued intracranial hypertension to maintain the intracranial pressure at less than 20 torr, provided that serum osmolality did not exceed 320 m0s/L."
Outcomes	Mortality at 3 and 12 months following injury, and uncontrolled ICP and mean of the worst CPP during treatment.
Notes	Four of the major teaching hospitals at the University of Toronto, Canada recruited people with TBI. The study recruited from April 1980 through October 1982.
Risk of bias
Bias	Authors' judgement	Support for judgement
Random sequence generation (selection bias)	Low risk	Quote: "...determined by a random number generator." p.436
Allocation concealment (selection bias)	Low risk	Quote: "Mechanism of randomisation was the opening of a serially numbered, sealed envelope taken from one of two packages of envelopes, one for patients who had had intracranial hematomas removed and one for those who developed raised intracranial pressure from brain injury alone." p.436 Quote: "The physician caring for the patient could not predict which drug was to be prescribed as initial treatment prior to opening the envelope." p.436
Blinding (performance bias and detection bias) Participants	Low risk	Review authors' judgement: Low risk as people had a severe head injury with a Glasgow Coma Scale score of less than 7 and therefore had reduced cognitive function.
Blinding (performance bias and detection bias) Treating physicians	High risk	Correspondence with the author in 2012. Quote: "... the treating physicians were aware of the patients' allocations...."
Blinding of outcome assessment (detection bias) All outcomes	High risk	Correspondence with the author in 2012. Quote: "The treatment records were reviewed which resulted in the reviewers knowing the allocation of the patient. Because the intracranial pressure measurements were numerical they were not subject to much debate."
Incomplete outcome data (attrition bias) All outcomes	Low risk	Complete outcome data were reported for all outcomes used in this review.
Selective reporting (reporting bias)	Unclear risk	Correspondence with the author in 2012. Quote: "The study protocol is long gone, but the experiment is described in the publication."
Other bias	Unclear risk	None known.