Background Most people who stop smoking gain weight, on average about 7kg in the long term. There are some interventions that have been specifically designed to tackle smoking cessation whilst also limiting weight gain. Many smoking cessation pharmacotherapies and other interventions may also limit weight gain. Objectives This review is divided into two parts. (1) Interventions designed specifically to aid smoking cessation and limit post-cessation weight gain (2) Interventions designed to aid smoking cessation that may also plausibly have an effect on weight Search strategy Part 1: We searched the Cochrane Tobacco Addiction Group's Specialized Register which includes trials indexed in MEDLINE, EMBASE, SciSearch and PsycINFO, and other reviews and conference abstracts. Part 2: We searched the included studies of Cochrane smoking cessation reviews of nicotine replacement therapy, antidepressants, nicotine receptor partial agonists, cannabinoid type 1 receptor antagonists (rimonabant), and exercise interventions, published in Issue 4, 2008 of The Cochrane Library. Selection criteria Part 1: We included trials of interventions designed specifically to address both smoking cessation and post-cessation weight gain that had measured weight at any follow-up point and/or smoking six months or more after quitting. Part 2: We included trials from the selected Cochrane reviews that could plausibly modify post-cessation weight gain if they had reported weight gain by trial arm at end of treatment or later. Data collection and analysis We extracted data in duplicate on smoking and weight for part 1 trials, and on weight only for part 2. Abstinence from smoking is expressed as a risk ratio (RR), using the most rigorous definition of abstinence available in each trial, and biochemically validated rates if available. The outcome is expressed as the difference in weight change between trial arms from baseline. Where appropriate, we performed meta-analysis using the Mantel-Haenszel method for smoking and inverse variance for weight using a fixed-effect model. Main results We found evidence that pharmacological interventions aimed at reducing post-cessation weight gain resulted in a significant reduction in weight gain at the end of treatment (dexfenfluramine (-2.50kg [-2.98kg to -2.02kg], fluoxetine (-0.80kg [-1.27kg to -0.33kg], phenylpropanolamine (PPA) (-0.50kg [-0.80kg to -0.20kg], naltrexone (-0.76kg [-1.51kg to -0.01kg])). No evidence of maintenance of the treatment effect was found at six or 12 months. Among the behavioural interventions, only weight control advice was associated with no reduction in weight gain and with a possible reduction in abstinence. Individualized programmes were associated with reduced weight gain at end of treatment and at 12 months (-2.58kg [-5.11kg to -0.05kg]), and with no effect on abstinence (RR 0.74 [0.39 to 1.43]). Very low calorie diets (-1.30kg (-3.49kg to 0.89kg] at 12 months) and cognitive behavioural therapy (CBT) (-5.20kg (-9.28kg to -1.12kg] at 12 months) were both associated with improved abstinence and reduced weight gain at end of treatment and at long-term follow up. Both bupropion (300mg/day) and fluoxetine (30mg and 60mg/day combined) were found to limit post-cessation weight gain at the end of treatment (-0.76kg [-1.17kg to -0.35kg] I-2= 48%) and -1.30kg [-1.91kg to -0.69kg]) respectively. There was no evidence that the weight reducing effect of bupropion was dose-dependent. The effect of bupropion at one year was smaller and confidence intervals included no effect (-0.38kg [-2.001kg to 1.24kg]). We found no evidence that exercise interventions significantly reduced post-cessation weight gain at end of treatment but evidence for an effect at 12 months (-2.07kg [-3.78kg, -0.36kg]). Treatment with NRT resulted in attenuation of post-cessation weight gain (-0.45kg [-0.70kg, - 0.20kg]) at the end of treatment, with no evidence that the effect differed for different forms of NRT. The estimated weight gain reduction was similar at 12 months (-0.42kg [-0.92kg, 0.08kg]) but the confidence intervals included no effect. There were no relevant data on the effect of rimonabant on weight gain. We found no evidence that varenicline significantly reduced post-cessation weight gain at end of treatment and no follow-up data are currently available. One study randomizing successful quitters to 12 more weeks of active treatment showed weight to be reduced by 0.71kg (-1.04kg to -0.38kg). In three studies, participants taking bupropion gained significantly less weight at the end of treatment than those on varenicline (-0.51kg [-0.93kg to -0.09kg]). Authors' conclusions Behavioual interventions of general advice only are not effective and may reduce abstinence. Individualized interventions, very low calorie diets, and CBT may be effective and not reduce abstinence. Exercise interventions are not associated with reduced weight gain at end of treatment, but may be associated with worthwhile reductions in weight gain in the long term, Bupropion, fluoxetine, nicotine replacement therapy, and probably varenicline all reduced weight gain while being used. Although this effect was not maintained one year after quitting for bupropion, fluoxetine, and nicotine replacement, the evidence is insufficient to exclude a modest long-term effect. The data are not sufficient to make strong clinical recommendations for effective programmes.
