A comparative study using data from the US Food and Drug Administration (FDA) has found that the risk of depression and suicidal/self-harm behaviour is substantially greater with varenicline than with other smoking cessation treatments. However, although the relative risk is striking, the absolute risk of harm is low.

Adverse event reports for the period 1998-2010 were identified for varenicline, bupropion and nicotine replacement products. Reports were also obtained for a comparison group of antibiotics.

Clearly, the number of adverse reports will reflect not only the relative risk but also the number of prescriptions issued for each drug or drug group; it may also have been the case that reporting rates were higher for newer drugs. Therefore, odds were used as the primary endpoint i.e. the number of cases of suicidal/self-harm behaviour or depression divided by the total number of other serious adverse events for that drug.

There were 3,249 reported cases of suicidal/self-harm behaviour or depression in individuals using smoking cessation therapy. Of these, 2,925 reports were for varenicline (90%), 229 for bupropion (7%) and 95 for nicotine replacement (3%). There were 295 completed suicides: 272 for varenicline (92%), 19 for bupropion (6%) and four for nicotine replacement (1%).

Compared with nicotine replacement, the odds ratio (OR) for varenicline was 8.4 (95% CI: 6.8-10.4). For bupropion, the OR was 2.9 (95% CI: 2.3-3.7). Compared with antibiotic treatment, the ORs were: varenicline 36.6, bupropion 12.5 and nicotine replacement 4.3 respectively.

The number of adverse event reports (AERs) for varenicline will have been substantially reduced by the fact that it was only available for the final four years of the study period. At the same time, the number of AERs is likely to have been substantially inflated by stimulated reporting: in the UK, there was a trebling of suicide-related AERs after regulatory warnings were issued in 2008.¹ Calculation of OR adjusts for the first, but not the second of these sources of bias.

The findings of this study corroborate the UK MHRA data (22 completed suicides with varenicline, six with bupropion and none with nicotine replacement).²

In contrast, data from the UK General Practice Database did not show an increased risk with varenicline, but the outcome of this study may have been influenced by under-reporting of adverse events to GPs.¹

The authors of the present study conclude that varenicline is ‘unsuitable for first-line use in smoking cessation.’ However, compared with non-smokers, smokers are twice as likely to become depressed,³ and more than twice as likely to commit suicide.⁴

Consequently, although smoking cessation in the short term is associated with an increased risk of depression5 and suicidal ideation, it may reduce the risk of suicide in the long term.⁴ There is evidence to suggest that varenicline is superior to both bupropion and NRT in achieving continuous abstinence.⁶ Therefore, it is more likely to reduce the long-term risk of suicide and this may offset the short-term risks. Moreover, although the relative risk is striking, the absolute risk of harm is low.

I think it is reasonable to continue to offer varenicline as a first-line treatment, provided:

• We counsel patients to stop their treatment and seek immediate advice if they develop agitation, depression or suicidal thoughts

• We closely monitor those with a previous history of psychiatric illness

• We do not prescribe to high risk patients e.g. those with a previous history of attempted suicide

• We liaise closely with smoking cessation services: I would advocate a shared-care protocol clarifying mutual responsibilities and the importance of passing on relevant information.

Dr Phillip Bland

• Moore TJ, Furberg CD, Glenmullen J et al. Suicidal behaviour and depression in smoking cessation treatments. PloS ONE 6(11): e27016.doi:10.1371.journal.pone.0027016