McCarthy 2008.

Methods	Open‐label behavioural intervention and placebo‐controlled pharmacotherapy intervention randomised clinical trial Study period: January 2001 to October 2003
Participants	N = 463 Participants were smokers seeking treatment recruited through mass media and screened by 3 rounds (telephone, group orientation, office visit). Inclusion criteria: aged 18 years or older who reported smoking ≥ 10 cigarettes/d and whose expired carbon monoxide (CO) levels exceeded 9 ppm. Participants reported being motivated to quit smoking (3 or 4 on a 4‐point self‐report scale) and being willing to fulfil study requirements. Exclusion criteria: serious psychopathology (bipolar disorder or psychosis), current depression, contraindications to use of bupropion SR (e.g. uncontrolled hypertension, history of seizure disorder, history of eating disorders, current heavy drinking, risk of pregnancy, current breastfeeding). Participants were excluded if their score on the Center for Epidemiologic Studies Depression (CES‐D) scale was above 16, except when an interview with a licensed clinician suggested that symptoms were related to a cause other than clinical depression.
Interventions	Active (bupropion SR) pharmacotherapy and individual targeted counseling (n = 113) Active pharmacotherapy, general counseling (N = 116) Placebo pharmacotherapy, individual targeted counseling (N = 121) Placebo pharmacotherapy, general counseling (N = 113) Participants attended 5 office visits in the 3 weeks before the quit date. Participants received 9 weeks of study pills (active or placebo), to begin 1 week before and to end 8 weeks after the planned quit date. Participants attended another 8 office visits over the 8 weeks following the quit date (provided breath samples at each visit and a blood sample at baseline and at end of treatment), then completed 10 monthly follow‐up calls. All participants received education regarding medication use and adherence, quit day information, and general encouragement, and completed electronic diaries for 2 weeks before and 4 weeks after the quit date. Participants randomised to individual targeted counseling received eight 10‐minute individual counseling sessions (2 prequit, 1 quit day, 5 postquit over 4 weeks).
Outcomes	Primary outcomes: electronic diary and retrospective report of lapse and of relapse (7 days) with 7‐day point prevalence abstinence confirmed via CO at each office visit and cotinine testing at end of treatment; 7‐day point prevalence abstinence at 6 and 12 months with CO testing Secondary outcomes: prolonged abstinence outcomes at end of treatment, at 6 and 12 months
Funding source	This work was supported by Transdisciplinary Tobacco Use Research Center grants CA084724 from the National Cancer Institute and DA19706 from the National Institute on Drug Abuse. GlaxoSmithKline provided complimentary active and placebo medication used in this study. GlaxoSmithKline was not involved in the design, data collection, analysis, or reporting of this study.
Declaration of interest	DEJ has received research support from Nabi Biopharmaceutical and Pfizer, Inc., and consulting fees from Nabi Biopharmaceutical. SS serves as consultant to GlaxoSmithKline Consumer Healthcare on an exclusive basis regarding over‐the‐counter smoking cessation products and is a partner in a company that is developing a new nicotine medication. He is a cofounder of invivodata, Inc., which provides electronic diary services for clinical research. In 1998 the University of Wisconsin appointed MCF to a named Chair made possible by an unrestricted gift to the university from GlaxoWellcome.
Notes	This work did not report analyses of pharmacogenetics, but these were reported in Bergen 2013.
Risk of bias
Bias	Authors' judgement	Support for judgement
Random sequence generation (selection bias)	Low risk	Randomisation via random number list
Allocation concealment (selection bias)	Unclear risk	Study pills did not differ in appearance between active and placebo but were packaged in containers before enrolment of participants. Randomisation was done via a random number list.
Blinding of participants and personnel (performance bias) All outcomes	Unclear risk	Research staff and participants were blind to pharmacotherapy randomisation but were not blind to counseling randomisation.
Blinding of outcome assessment (detection bias) All outcomes	Low risk	Outcome assessment was done by electronic diary through end of treatment, and by interview (self‐report), performed in office or by phone calls, with biochemical abstinence verified at each office visit (CO) and at 6 and 12 months (cotinine). Discordance between reports or verification methods analysed in multiple ways, but final prevalences highly similar
Incomplete outcome data (attrition bias) All outcomes	Low risk	Attrition was substantial and did not differ by treatment arm at quit, end of treatment, or any other follow‐up point. Attrition was related to various covariates included in abstinence models in the primary paper.
Selective reporting (reporting bias)	Low risk	Primary and secondary outcomes were reported.
Other bias	High risk	DNA samples were provided for 41% of self‐identified whites for analysis of pharmacogenetics. Most (˜ 99%) DNA samples were successfully genotyped. Analyses of pharmacogenetics in Bergen 2013 are distributed equally among treatment groups, so selection bias is unlikely in this sample.