Gooday 1995

Methods	Double‐blind, single‐centre, placebo‐controlled RCT. Cross‐over design: all participants were crossed over to alternative intervention.
Participants	10 participants with cirrhosis and subclinical (corresponding to minimal) hepatic encephalopathy diagnosed based on a score on the Digit Symbol Substitution test of < 1 SD of the age‐matched normative mean. Mean age ± SD: 53.9 ± 7.4 years. Proportion of men: 80%. Aetiology of cirrhosis: alcohol 60%; hepatitis B/C 20%. Proportion testing positive for benzodiazepines at baseline (Table 7): apparently not performed (not specifically stated).
Interventions	Intervention comparison: intravenous infusion flumazenil 0.2 mg over an unspecified time versus placebo (saline). Total dose of flumazenil: 0.2 mg. Washout period: 1 week. Cointerventions: none described.
Outcomes	Outcomes included in meta‐analyses: mortality and serious adverse events (Table 8) assessed at end of the intervention.
Neuropsychiatric assessment	At baseline: Digit Symbol Substitution Test. At baseline and post infusion: Simple reaction time; Complex reaction time; Auditory Verbal Learning Test; Digit Symbol Substitution Test; Digits forward and backwards. Duration of follow‐up and timing of tests not described.
Inclusion period (date)	Not described.
Country	UK.
Notes	Included data: we received additional (unpublished) information about the trial methods and number of participants allocated to flumazenil/placebo during the first allocation period via email in 2003 when conducting the previous version of this review. The trial did not evaluate the number of participants with an overall improvement in hepatic encephalopathy. Therefore, we were unable to include the trial in our analyses of this outcome measure.
Risk of bias
Bias	Authors' judgement	Support for judgement
Random sequence generation (selection bias)	Low risk	Random number table.
Allocation concealment (selection bias)	Low risk	Serially numbered opaque sealed envelopes used in administration of concealed drug containers.
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.
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	Unclear risk	No information provided.
Other bias	High risk	Primary investigators described a significant drug by order and group by drug by order interaction.
Overall assessment	High risk	High risk of bias.