Wright 2006.

Methods	Block‐randomised, double‐blind, placebo‐controlled phase II trial
Participants	Adults with acute severe TBI and a GCS score 4‐12 after resuscitation and stabilisation within 11 hours of injury Each participant was assigned to 1 of 8 clinical subgroups defined by sex, race (black versus others), and TBI severity (index GCS scores 4 to 8 were categorised as severe; 9 to 12 as moderate). Within each subgroup, permuted block randomisation assigned 4 of every 5 consecutive participants to progesterone and the other to placebo. A 4:1 randomisation scheme was used to increase the number of participants receiving progesterone while maintaining blinding. Exclusion criteria: indeterminate time of injury; pregnancy; a family reported history of active cancer, acute stroke or of older stroke with residual motor deficits; acute or chronic spinal cord injury with neurologic deficits; a blood alcohol concentration > 250 mg/dL; penetrating brain injury; < 18 years old
Interventions	Intervention group: progesterone was mixed in Intralipid 20% at a concentration designed to deliver a loading dose of 0.71 mg/kg at 14 mL/h for the first hour when a participant was enrolled. Then the infusion was reduced to 10 mL/h to deliver 0.5 mg/kg/h for the next 11 hours. Five additional 12‐hour maintenance infusions were delivered at the standard rate of 10 mL/h, for a total of 3 days of treatment. Control group: placebo
Outcomes	Mortality; dichotomised GOS; DRS; duration of coma; duration of post‐traumatic amnesia in 2 subgroups (index GCS scores 4‐8 severe; 9‐12 moderate) at 30 days postinjury, ICP, body temperature, blood pressure during the first 3 days of treatment and for 1 day after treatment, adverse events.
Notes	Funding and support: "Supported by a grant from the National Institute for Neurological Disorders and Stroke, National Institutes of Health (1 R01 NS‐39097‐01A1 to AK) and the General Clinical Research Center at Emory University and Grady Memorial Hospital".
Risk of bias
Bias	Authors' judgement	Support for judgement
Random sequence generation (selection bias)	Low risk	Quote: "permuted block randomisation assigned 4 of every 5 consecutive patients to progesterone and the other to placebo". (p 393) Comment: probably done
Allocation concealment (selection bias)	Low risk	Quote: "permuted block randomisation" (p 393) Quote: "Drug kits were prepared and randomised off site by Emory's Investigational Drug Center. These kits were indistinguishable with respect to treatment assignment".(p 393) Comment: probably done
Blinding (performance bias and detection bias) Subjective outcome	Low risk	Quote: "Double blind" "Drug kits were prepared and randomised off site by Emory's Investigational Drug Center. These kits were indistinguishable with respect to treatment assignment". (p 393) Comment: probably done
Blinding (performance bias and detection bias) Mortality	Low risk	Obtained from medical records; review authors do not believe this introduced bias.
Incomplete outcome data (attrition bias) Mortality	Low risk	100 participants were randomised; 1 participant who was randomised to progesterone died before the infusion could be started (this subject's data were retained in the analysis under the principle of ITT). Treatment for 3 participants was discontinued (1 was taken into police custody, 2 died during infusion). None of the participants was lost to follow‐up at 30 days.
Incomplete outcome data (attrition bias) Favorable outcome	Low risk	92 participants were contacted to assess their functional status. All the data are presented.
Selective reporting (reporting bias)	Low risk	The protocol for this trial was presented in Clinicaltrials.gov and any remaining information was obtained from study authors. It was clear that the published report included all expected outcomes.
Other bias	Low risk	‐‐