Spring 2007.

Methods	FLUOXETINE Randomized controlled trial Setting: clinic, USA Recruitment: community volunteers
Participants	247 smokers, >= 10 CPD; 54% F, av. age 44, av. CPD 23, 44% history of MDD
Interventions	1. Fluoxetine 60 mg (titrated up over 2 w) for 12 weeks 2. Placebo Both arms: group behavioural counselling, 9 meetings over 12 weeks
Outcomes	Abstinence at 6m (prolonged from 2 w after quit date) Validation: CO < 10 ppm, urine cotinine < 20 ng/ml
Notes	First included as Spring 2004 with unpublished data. Full publication reports sustained abstinence Funding: National Institutes of Health, Veterans Affairs. Medication provided by Eli Lilly and Company.
Risk of bias
Bias	Authors' judgement	Support for judgement
Random sequence generation (selection bias)	Low risk	"The study pharmacist stratified participants by depression history and used computer‐ generated random numbers to assign them to drug or placebo."
Allocation concealment (selection bias)	Unclear risk	Allocated by unblinded pharmacist, method not described
Blinding (performance bias and detection bias) All outcomes	Unclear risk	Double blind, "Research staff and participants were blinded to medication status." "Drug assignment was guessed correctly by 59.8% of placebo and 64.6% of fluoxetine participants. Facilitators guessed correctly for 65.3% of placebo and 55.6% of fluoxetine participants."
Incomplete outcome data (attrition bias) All outcomes	Low risk	Withdrawals/lost to follow‐up 40% for fluoxetine, 48% placebo. Authors report similar results from missing assumed smoking and GEE analyses. All participants included in MA.