Pomier‐Layrargues 1994.

Methods	Double‐blind, single‐centre, cross‐over, placebo‐controlled RCT. Cross‐over design: investigators only crossed over participants who remained in Grade IV coma, 24 hours after the first study period.
Participants	21 participants with cirrhosis and acute hepatic encephalopathy (Grade IV). Precipitating factors are described (Table 5). Mean age ± SD: flumazenil: 52.7 ± 5.4 years; placebo: 57.4 ± 9.0 years. Proportion of men: 81%. Aetiology of cirrhosis: alcohol 62%; hepatitis B/C 5%. Proportion testing positive for benzodiazepines at baseline (Table 6): 4/21 (19%) participants.
Interventions	Intervention comparison: intravenous infusion flumazenil 2 mg over 5 minutes versus placebo (saline). Total dose of flumazenil: 2 mg. Washout period: 24 hours. Cointerventions: lactulose 30 mL 4 times daily.
Outcomes	Outcomes included in meta‐analyses: mortality, hepatic encephalopathy (Table 2), and serious adverse events (Table 7) assessed after a maximum follow‐up of 24 hours.
Neuropsychiatric assessment	Baseline and post infusion: Modified Glasgow Coma Scale (Table 3) undertaken at baseline and every 15 minutes for up to 5.5 hours post infusion; Continuous electroencephalography (Table 4) 15 minutes before and 15 minutes after the infusion.
Inclusion period (date)	March 1988 to February 1992.
Country	Canada.
Notes	Included data: we only included data from the first treatment period in our analyses.
Risk of bias
Bias	Authors' judgement	Support for judgement
Random sequence generation (selection bias)	Low risk	Random numbers.
Allocation concealment (selection bias)	Low risk	Blinded administration of flumazenil or placebo.
Blinding of participants and personnel (performance bias) All outcomes	Low risk	Blinding of participants and personnel.
Blinding of outcome assessment (detection bias) All outcomes	Low risk	Blinding of outcome assessment.
Incomplete outcome data (attrition bias) All outcomes	Low risk	No missing outcome data and all participants included in analyses.
Selective reporting (reporting bias)	Low risk	Trial described clinically relevant outcomes. We had no access to information about outcomes described in the original protocol or information in trial registries.
For‐profit funding	High risk	Technical assistance from Hoffmann‐La Roche Ltd., Canada and Nutley, NJ, USA.
Other bias	Low risk	No other biases.
Overall assessment	High risk	High risk of bias.