Background and review questions
Some medicines and supplements that have been used to treat depression (antidepressants) have also been tested to see whether they can help people to stop smoking. Two of these treatments - bupropion (sometimes called Zyban) and nortriptyline - are sometimes given to help people quit smoking. This review looks at whether using antidepressants actually helps people to stop smoking (for six months or longer), and also looks at the safety of using these medicines.
This review includes 115 studies looking at how helpful and safe different antidepressants are when used to quit smoking. Most of the studies were conducted in adults. We included studies of any length when looking at safety, but studies needed to be at least six months long when assessing whether people had managed to quit smoking. The evidence is up to date to May 2019.
Using the antidepressant, bupropion, makes it 52% to 77% more likely that a person will successfully stop smoking, which is equal to five to seven more people successfully quitting for six months or more for every one hundred people who try to quit. There is evidence that people who use the antidepressant, nortriptyline, to quit smoking also improve their chances of success. There is not enough evidence to determine whether other antidepressants help people to quit smoking.
There is evidence that bupropion increases unwanted effects, particularly those relating to mental health, and that unwanted effects may increase the chance that people stop using the medicine. However, the evidence does not suggest that bupropion is more likely to result in death, hospitalization, or life-threatening events, like seizures. There is not enough information to draw clear conclusions about the safety of nortriptyline for stopping smoking.
The evidence does not suggest that taking bupropion at the same time as other stop-smoking medicines, like varenicline (sometimes known as Champix or Chantix) or nicotine replacement therapy makes people more likely to quit smoking. People are as likely to quit smoking when using bupropion as when using nortriptyline or nicotine replacement therapy, however people using varenicline are more likely to quit than those using bupropion.
Certainty of evidence
There is high-certainty evidence that bupropion helps people to quit smoking, meaning further research is very unlikely to change this conclusion. However, there is also high-certainty evidence to suggest that people using bupropion are more likely to stop taking the medicine because of unpleasant effects than those taking a pill without medication (a placebo). The certainty of the evidence was moderate, low or very low for the other key questions we looked at. This means that the findings of those questions may change when more research is carried out. In most cases this was because there were not enough studies or studies were too small.
There is high-certainty evidence that bupropion can aid long-term smoking cessation. However, bupropion also increases the number of adverse events, including psychiatric AEs, and there is high-certainty evidence that people taking bupropion are more likely to discontinue treatment compared with placebo. However, there is no clear evidence to suggest whether people taking bupropion experience more or fewer SAEs than those taking placebo (moderate certainty). Nortriptyline also appears to have a beneficial effect on smoking quit rates relative to placebo. Evidence suggests that bupropion may be as successful as NRT and nortriptyline in helping people to quit smoking, but that it is less effective than varenicline. There is insufficient evidence to determine whether the other antidepressants tested, such as SSRIs, aid smoking cessation, and when looking at safety and tolerance outcomes, in most cases, paucity of data made it difficult to draw conclusions. Due to the high-certainty evidence, further studies investigating the efficacy of bupropion versus placebo are unlikely to change our interpretation of the effect, providing no clear justification for pursuing bupropion for smoking cessation over front-line smoking cessation aids already available. However, it is important that where studies of antidepressants for smoking cessation are carried out they measure and report safety and tolerability clearly.
Whilst the pharmacological profiles and mechanisms of antidepressants are varied, there are common reasons why they might help people to stop smoking tobacco. Firstly, nicotine withdrawal may produce depressive symptoms and antidepressants may relieve these. Additionally, some antidepressants may have a specific effect on neural pathways or receptors that underlie nicotine addiction.
To assess the evidence for the efficacy, safety and tolerability of medications with antidepressant properties in assisting long-term tobacco smoking cessation in people who smoke cigarettes.
We searched the Cochrane Tobacco Addiction Specialized Register, which includes reports of trials indexed in the Cochrane Central Register of Controlled Trials (CENTRAL), MEDLINE, Embase, and PsycINFO, clinicaltrials.gov, the ICTRP, and other reviews and meeting abstracts, in May 2019.
We included randomized controlled trials (RCTs) that recruited smokers, and compared antidepressant medications with placebo or no treatment, an alternative pharmacotherapy, or the same medication used in a different way. We excluded trials with less than six months follow-up from efficacy analyses. We included trials with any follow-up length in safety analyses.
We extracted data and assessed risk of bias using standard Cochrane methods. We also used GRADE to assess the certainty of the evidence.
The primary outcome measure was smoking cessation after at least six months follow-up, expressed as a risk ratio (RR) and 95% confidence intervals (CIs). We used the most rigorous definition of abstinence available in each trial, and biochemically validated rates if available. Where appropriate, we performed meta-analysis using a fixed-effect model.
Similarly, we presented incidence of safety and tolerance outcomes, including adverse events (AEs), serious adverse events (SAEs), psychiatric AEs, seizures, overdoses, suicide attempts, death by suicide, all-cause mortality, and trial dropout due to drug, as RRs (95% CIs).
We included 115 studies (33 new to this update) in this review; most recruited adult participants from the community or from smoking cessation clinics. We judged 28 of the studies to be at high risk of bias; however, restricting analyses only to studies at low or unclear risk did not change clinical interpretation of the results. There was high-certainty evidence that bupropion increased long-term smoking cessation rates (RR 1.64, 95% CI 1.52 to 1.77; I2 = 15%; 45 studies, 17,866 participants). There was insufficient evidence to establish whether participants taking bupropion were more likely to report SAEs compared to those taking placebo. Results were imprecise and CIs encompassed no difference (RR 1.16, 95% CI 0.90 to 1.48; I2 = 0%; 21 studies, 10,625 participants; moderate-certainty evidence, downgraded one level due to imprecision). We found high-certainty evidence that use of bupropion resulted in more trial dropouts due to adverse events of the drug than placebo (RR 1.37, 95% CI 1.21 to 1.56; I2 = 19%; 25 studies, 12,340 participants). Participants randomized to bupropion were also more likely to report psychiatric AEs compared with those randomized to placebo (RR 1.25, 95% CI 1.15 to 1.37; I2 = 15%; 6 studies, 4439 participants).
We also looked at the safety and efficacy of bupropion when combined with other non-antidepressant smoking cessation therapies. There was insufficient evidence to establish whether combination bupropion and nicotine replacement therapy (NRT) resulted in superior quit rates to NRT alone (RR 1.19, 95% CI 0.94 to 1.51; I2 = 52%; 12 studies, 3487 participants), or whether combination bupropion and varenicline resulted in superior quit rates to varenicline alone (RR 1.21, 95% CI 0.95 to 1.55; I2 = 15%; 3 studies, 1057 participants). We judged the certainty of evidence to be low and moderate, respectively; in both cases due to imprecision, and also due to inconsistency in the former. Safety data were sparse for these comparisons, making it difficult to draw clear conclusions.
A meta-analysis of six studies provided evidence that bupropion resulted in inferior smoking cessation rates to varenicline (RR 0.71, 95% CI 0.64 to 0.79; I2 = 0%; 6 studies, 6286 participants), whilst there was no evidence of a difference in efficacy between bupropion and NRT (RR 0.99, 95% CI 0.91 to 1.09; I2 = 18%; 10 studies, 8230 participants).
We also found some evidence that nortriptyline aided smoking cessation when compared with placebo (RR 2.03, 95% CI 1.48 to 2.78; I2 = 16%; 6 studies, 975 participants), whilst there was insufficient evidence to determine whether bupropion or nortriptyline were more effective when compared with one another (RR 1.30 (favouring bupropion), 95% CI 0.93 to 1.82; I2 = 0%; 3 studies, 417 participants). There was no evidence that any of the other antidepressants tested (including St John's Wort, selective serotonin reuptake inhibitors (SSRIs), monoamine oxidase inhibitors (MAOIs)) had a beneficial effect on smoking cessation. Findings were sparse and inconsistent as to whether antidepressants, primarily bupropion and nortriptyline, had a particular benefit for people with current or previous depression.