Hall 2008.

Methods	Randomised parallel‐group open‐label efficacy clinical trial consisting of 2 phases: phase 1 lasting 12 weeks, during which all participants receive a standard treatment of NRT, bupropion, and 5 group counseling sessions; followed by phase 2, randomisation to 1 of 5 treatments for another 40 weeks. Outcomes are measured at multiple time points to 104 weeks. Study period: 2002 to 2004
Participants	N = 407 Recruited through multi‐media advertising, public service announcements, flyers, and direct mailing Inclusion criteria: treatment‐seeking smokers 18 years of age or older who currently smoke 10 cigarettes/d, report a regular smoking history ≥ 5 years, and answer "yes" to the question, “Do you smoke within 30 minutes of arising?” Exclusion criteria: history of seizure or head injury resulting in unconsciousness; any condition that might predispose to seizures (brain tumour or stroke); current or history of anorexia nervosa or bulimia; any disease acutely life‐threatening or so severe that the patient is judged unable to comply with the protocol; use of a protease inhibitor or MAO inhibitor within the past 2 weeks; current use of psychiatric drugs that would interfere with interpretation of study results, including antidepressants; treatment for alcohol dependence during the past year, or evidence of alcohol abuse so severe that the patient is judged potentially unable to comply with the protocol; patients who know they are leaving the Bay Area within the study period and non‐English speakers; suicidal or homicidal ideation; current major depression; history of bipolar disorder; recent (within 12 months) myocardial infarction; any other medical condition that would contraindicate use of NRT or bupropion; physical limitation so severe that participation in a programme of moderate exercise is not possible; and pregnancy or lactation
Interventions	All participants receive 10 weeks of NRT (patch, tapering from 21 mg to 7 mg over 8 weeks, starting at week 3) and 12 weeks of bupropion treatment and 5 mandatory group counseling sessions. At week 11, subjects are randomly assigned to 1 of 5 treatment groups: (a) no further treatment, with assessments at weeks 12, 24, 36, 52, 64, and 104 weeks (standard assessments for all treatments); (b) extended active medication (bupropion) treatment through week 52 with low‐intensity (monthly) counseling with medical staff; (c) extended placebo medication treatment through week 52 with low‐intensity counseling with medical staff; (d) extended active medication treatment with high‐intensity (session 20 to 40 minutes in duration at weeks 12, 14, 16, 18, 20, 24, 28, 32, 36, 44, and 52, and telephone contact at weeks 13, 15, 18, 22, 26, 30, 34, 36, 40, and 48) counseling; and (e) extended placebo medication treatment with high‐intensity counseling. High‐intensity counseling sessions included additional information focusing on motivation, social support, mood management, weight gain, and dependence/withdrawal.
Outcomes	Primary outcome: biochemically verified 7‐day point prevalence abstinence at weeks 12, 24, 52, 64, and 104
Funding source	The clinical trial was supported by a grant from the National Institute on Drug Abuse (R01DA015732, Maintaining Abstinence in Chronic Smokers, PI: Sharon M Hall).
Declaration of interest	Declaration of interest was not reported.
Notes	Analyses of pharmacogenetics were described in Bergen 2013.
Risk of bias
Bias	Authors' judgement	Support for judgement
Random sequence generation (selection bias)	Low risk	Computerised allocation was performed by study statistician who did not have contact with participants.
Allocation concealment (selection bias)	Low risk	Assignment for individual participants was transmitted electronically to staff.
Blinding of participants and personnel (performance bias) All outcomes	Unclear risk	The trial was open‐label, but other aspects of the trial suggest rigorous management and lower probability of high performance bias to unclear.
Blinding of outcome assessment (detection bias) All outcomes	Low risk	Outcome assessment was done by self‐report and by 2 biochemical methods using exhaled breath and urine (i.e. very thorough outcome assessment).
Incomplete outcome data (attrition bias) All outcomes	Low risk	Attrition rates were low for the companion trial.
Selective reporting (reporting bias)	Low risk	Outcome assessment was very thorough, with 3 pieces of information required to note an individual as abstinent. In the companion trial, only 17/905 urine determinations had data discordant with outcomes of self‐report and CO measurement (< 2%).
Other bias	Low risk	Analyses of pharmacogenetics were performed on ˜ 37% of total study, ˜ 50% of self‐identified “Caucasian”. Minimum arm N was 26. Data show no significant differences by arm with respect to proportion of the arm genotyped. Not generalisable to other ancestries