Furieri 2007.

Methods	Randomised placebo‐controlled double‐blind study
Participants	60 participants; mean age 44.2 years; 100% male; 6.6 years on average of education; 37% married; 17% employed; 16 years on average age of alcohol use onset; 27 years of drinking on average; 17 drinks per day on average in the past 90 days; CIWA‐Ar score 25 on average Inclusion criteria: 18 to 65 years old; consuming at least 35 drinks per week during the past year and the past 90 days; being abstinent from alcohol from no longer than 14 days before baseline; meeting DSM‐IV criteria for alcohol dependence (First 1994); being in stable clinical condition; having normal serum liver transaminases; having a plasma GGT level less than 800 U/L; being diagnosed as non‐demented with Mini Mental State Examination (Folstein 1975); having no severe withdrawal signs or symptoms; scoring less than 15 on the CIWA‐Ar (Sullivan 1989); having completed seven‐day treatment for acute alcohol withdrawal Exclusion criteria: meeting diagnostic criteria for other substance intoxication or withdrawal or unstable medical or mental disorder other than alcohol dependence, except nicotine and/or caffeine; having convulsions or delirium during abstinence from alcohol; having used pharmacological agents known to reduce the convulsive threshold or to alter alcohol withdrawal or craving during the past 30 days; having a previous history of drug hypersensitivity or adverse reactions to gabapentin, diazepam or other benzodiazepines and haloperidol
Interventions	(1) Gabapentin, 30 participants; (2) placebo, 30 participants Drug dose: gabapentin up to 600 mg/d Participants involved in weekly BBCET (Johnson 2003b) Setting: outpatient Duration: four weeks. Country of origin: Brazil
Outcomes	Primary outcomes: alcohol use with the TLFB (Sobell 1992) as follows: drinks per day, drinks per drinking day, vital signs, adverse events, percentage of heavy drinking days, percentage of days of abstinence; alcohol craving using the OCDS (Bohn 1996): haematological and biochemical measurements, including transaminases and plasma GGT
Notes	All participants received one‐week detoxification before initiation of double‐blind treatment
Risk of bias
Bias	Authors' judgement	Support for judgement
Random sequence generation (selection bias)	Low risk	Randomisation carried out by a computer programme
Allocation concealment (selection bias)	Low risk	Quote: "The randomization code was held by a research supervisor, to be broken only in case of emergency". Study medication dispensed in medication containers, each labelled with participant code
Incomplete outcome data (attrition bias) All outcomes	High risk	Only completers analysis performed
Blinding of participants and personnel (performance bias) all outcomes	Low risk	Double‐blind stated. Medication and matching placebo prepared by pharmaceutical company. Participants and psychiatrist blind to treatment condition
Blinding of outcome assessment (detection bias) subjective	Low risk	Investigator and supervisor blind to treatment condition
Blinding of outcome assessment (detection bias) objective	Low risk	Investigator and supervisor blind to treatment condition