What is the best way to avoid putting on weight after stopping smoking?
Key messages
We are not certain which programmes or treatments work best to help people avoid gaining weight in the long term (up to 12 months) when stopping smoking, or how they affect success in stopping smoking. This is because the evidence shows varied and unclear effects on weight gain. Further studies should continue to look at how to limit weight gain in people who are stopping smoking. Future studies of new medicines to help people stop smoking should also measure changes in their weight.
Stopping smoking and weight gain
If you smoke, the best thing you can do for your health is to stop. But people often put on weight when they stop smoking, usually in the first few months of stopping. Gaining weight might undermine some of the benefits of stopping smoking, and might also affect the motivation of some people trying to stop smoking.
What did we want to find out?
Some programmes to help people stop smoking specifically target weight control. Other ways to help people stop smoking might also affect their weight; these include: exercise programmes, taking medicines, and using nicotine replacement therapy (NRT).
We wanted to find out the best ways to stop weight gain when stopping smoking.
What did we do?
In this update of a previously-published review, we searched for studies that tested:
- specific programmes for weight control while stopping smoking;
- other ways to help people stop smoking, if those studies also measured weight change.
We were interested in:
- how many people stopped smoking for six months or 12 months;
- people's weight at the end of treatment, then after six months and after 12 months.
What did we find?
We found 116 studies in total:
37 studies of specific programmes for weight control for people stopping smoking (21 new studies for this update); and 83 studies of other ways to help people stop smoking (27 new studies for this update). Four of these studies contributed to both.
The 37 studies of specific programmes tested behavioural programmes, including dieting, to manage weight in 11,514 people trying to stop smoking. Some of the behavioural programmes were acceptance-based, in which people also learn self-regulation skills (for example, how to deal with cravings) to help them keep to the behaviours needed to lose weight. Most studies (27) were done in the USA and others took place in Australia, Canada, China and Europe.
The 83 studies of other ways of stopping smoking included 46,248 people and looked at:
exercise programmes; taking NRT; taking a medicine called varenicline (used to help people stop smoking); or taking a medicine called fluoxetine (used to treat depression).
Of these studies, 39 were done in the USA and the rest took place in other countries around the world. Very few studies reported on side effects.
What are the main results of our review?
Programmes aimed at limiting weight gain
Compared with no programme or brief advice only, a personalized weight-management programme may reduce weight gain at the end of treatment, after six months and after 12 months. However, a weight-management programme without personalized assessment, planning and feedback may not reduce weight gain, and may reduce the number of people who stop smoking.
Compared with no programme, acceptance-based programmes for weight:
may help more people to stop smoking after six months and 12 months; but may make little to no difference to their weight gain.
Other programmes and treatments that might affect weight
Taking part in an exercise programme to help stop smoking may reduce weight gain after 12 months, compared with not taking part in one.
Using NRT probably reduces weight gain slightly after 12 months, compared with not using NRT.
Taking varenicline makes little difference to weight gain at the end of treatment, and may make little difference after six months or after 12 months.
Taking fluoxetine may reduce weight gain at the end of treatment, but we do not know how it affects weight gain after six months or 12 months.
What are the limitations of the evidence?
We are certain that there is no difference in weight gain at the end of treatment with varenicline, and further studies are unlikely to change this result. However, our confidence in all of the other evidence is limited, mainly because of small numbers of studies that could be compared, and small numbers of people taking part in them. The results varied widely, and there were not enough studies for us to be sure of the results. Our confidence is likely to change if further evidence becomes available.
How up to date is this evidence?
The evidence is current up to October 2020.
Overall, there is no intervention for which there is moderate certainty of a clinically useful effect on long-term weight gain. There is also no moderate- or high-certainty evidence that interventions designed to limit weight gain reduce the chances of people achieving abstinence from smoking.
Most people who stop smoking gain weight. This can discourage some people from making a quit attempt and risks offsetting some, but not all, of the health advantages of quitting. Interventions to prevent weight gain could improve health outcomes, but there is a concern that they may undermine quitting.
To systematically review the effects of: (1) interventions targeting post-cessation weight gain on weight change and smoking cessation (referred to as 'Part 1') and (2) interventions designed to aid smoking cessation that plausibly affect post-cessation weight gain (referred to as 'Part 2').
Part 1 - We searched the Cochrane Tobacco Addiction Group's Specialized Register and CENTRAL; latest search 16 October 2020.
Part 2 - We searched included studies in the following 'parent' Cochrane reviews: nicotine replacement therapy (NRT), antidepressants, nicotine receptor partial agonists, e-cigarettes, and exercise interventions for smoking cessation published in Issue 10, 2020 of the Cochrane Library. We updated register searches for the review of nicotine receptor partial agonists.
Part 1 - trials of interventions that targeted post-cessation weight gain and had measured weight at any follow-up point or smoking cessation, or both, six or more months after quit day.
Part 2 - trials included in the selected parent Cochrane reviews reporting weight change at any time point.
Screening and data extraction followed standard Cochrane methods. Change in weight was expressed as difference in weight change from baseline to follow-up between trial arms and was reported only in people abstinent from smoking. Abstinence from smoking was expressed as a risk ratio (RR). Where appropriate, we performed meta-analysis using the inverse variance method for weight, and Mantel-Haenszel method for smoking.
Part 1: We include 37 completed studies; 21 are new to this update. We judged five studies to be at low risk of bias, 17 to be at unclear risk and the remainder at high risk.
An intermittent very low calorie diet (VLCD) comprising full meal replacement provided free of charge and accompanied by intensive dietitian support significantly reduced weight gain at end of treatment compared with education on how to avoid weight gain (mean difference (MD) −3.70 kg, 95% confidence interval (CI) −4.82 to −2.58; 1 study, 121 participants), but there was no evidence of benefit at 12 months (MD −1.30 kg, 95% CI −3.49 to 0.89; 1 study, 62 participants). The VLCD increased the chances of abstinence at 12 months (RR 1.73, 95% CI 1.10 to 2.73; 1 study, 287 participants). However, a second study found that no-one completed the VLCD intervention or achieved abstinence.
Interventions aimed at increasing acceptance of weight gain reported mixed effects at end of treatment, 6 months and 12 months with confidence intervals including both increases and decreases in weight gain compared with no advice or health education. Due to high heterogeneity, we did not combine the data. These interventions increased quit rates at 6 months (RR 1.42, 95% CI 1.03 to 1.96; 4 studies, 619 participants; I2 = 21%), but there was no evidence at 12 months (RR 1.25, 95% CI 0.76 to 2.06; 2 studies, 496 participants; I2 = 26%).
Some pharmacological interventions tested for limiting post-cessation weight gain (PCWG) reduced weight gain at the end of treatment (dexfenfluramine, phenylpropanolamine, naltrexone). The effects of ephedrine and caffeine combined, lorcaserin, and chromium were too imprecise to give useful estimates of treatment effects. There was very low-certainty evidence that personalized weight management support reduced weight gain at end of treatment (MD −1.11 kg, 95% CI −1.93 to −0.29; 3 studies, 121 participants; I2 = 0%), but no evidence in the longer-term 12 months (MD −0.44 kg, 95% CI −2.34 to 1.46; 4 studies, 530 participants; I2 = 41%). There was low to very low-certainty evidence that detailed weight management education without personalized assessment, planning and feedback did not reduce weight gain and may have reduced smoking cessation rates (12 months: MD −0.21 kg, 95% CI −2.28 to 1.86; 2 studies, 61 participants; I2 = 0%; RR for smoking cessation 0.66, 95% CI 0.48 to 0.90; 2 studies, 522 participants; I2 = 0%).
Part 2: We include 83 completed studies, 27 of which are new to this update.
There was low certainty that exercise interventions led to minimal or no weight reduction compared with standard care at end of treatment (MD −0.25 kg, 95% CI −0.78 to 0.29; 4 studies, 404 participants; I2 = 0%). However, weight was reduced at 12 months (MD −2.07 kg, 95% CI −3.78 to −0.36; 3 studies, 182 participants; I2 = 0%).
Both bupropion and fluoxetine limited weight gain at end of treatment (bupropion MD −1.01 kg, 95% CI −1.35 to −0.67; 10 studies, 1098 participants; I2 = 3%); (fluoxetine MD −1.01 kg, 95% CI −1.49 to −0.53; 2 studies, 144 participants; I2 = 38%; low- and very low-certainty evidence, respectively). There was no evidence of benefit at 12 months for bupropion, but estimates were imprecise (bupropion MD −0.26 kg, 95% CI −1.31 to 0.78; 7 studies, 471 participants; I2 = 0%). No studies of fluoxetine provided data at 12 months.
There was moderate-certainty that NRT reduced weight at end of treatment (MD −0.52 kg, 95% CI −0.99 to −0.05; 21 studies, 2784 participants; I2 = 81%) and moderate-certainty that the effect may be similar at 12 months (MD −0.37 kg, 95% CI −0.86 to 0.11; 17 studies, 1463 participants; I2 = 0%), although the estimates are too imprecise to assess long-term benefit.
There was mixed evidence of the effect of varenicline on weight, with high-certainty evidence that weight change was very modestly lower at the end of treatment (MD −0.23 kg, 95% CI −0.53 to 0.06; 14 studies, 2566 participants; I2 = 32%); a low-certainty estimate gave an imprecise estimate of higher weight at 12 months (MD 1.05 kg, 95% CI −0.58 to 2.69; 3 studies, 237 participants; I2 = 0%).