Varenicline (aka Champix/Chantrix, a partial agonist of the [α4β2] nicotinic acetylcholine receptor) is a drug treatment for tobacco dependence. It is widely prescribed and works through reducing craving, withdrawal symptoms and the rewarding and reinforcing effects of tobacco.

A Cochrane review (Cahill et al., 2013), found that varenicline increased rates of successfully quitting more than two-fold relative to placebo; was more successful than bupropion; and did not differ significantly from nicotine patch treatment. You can read the Mental Elf blog on this review here.

However, based on case reports and post-marketing surveys, there have been associations drawn between varenicline’s use and increases in suicidality, depression, violence, psychosis, and an increased incidence of traffic accidents, resulting in prohibition within certain transportation industries.

Importantly, these associations are not supported by the existing randomised controlled trials and observational data. Discrepancies across evidence sources may be due to: study design, reporting bias, small sample size of RCTs, and various confounds, for example, those with pre-existing mental health disorders often smoke or the impact of the nicotine withdrawal syndrome.

Methods

• From the population of Sweden, 69,757 people (43,861 women) over 15 years of age were identified who were prescribed varenicline between Nov 2006 – Dec 2009. This left 7,847,679 people as the non-treated cohort.
• Treatment was defined as at least 1 collected prescription of varenicline, and treatment periods were 12 weeks (standard length). Therefore, any number of prescriptions collected within the first 12 weeks was considered 1 treatment period, any collected prescriptions collected after 12 weeks were considered a subsequent treatment period.

Outcomes

• Crime: all penal code offences, except traffic offences (source: National Crime Register)
• Psychiatric conditions: the incidence of new psychiatric conditions were taken and split into three categories: anxiety conditions, mood conditions, psychoses (Patient Register).
• Suicidal behaviour: Suicides/attempted suicides were those that involved death and/or emergency inpatient or outpatient hospital visits (Patient Register and Cause of Death Register)
• Substance abuse: alcohol use disorders and dependence, drug misuse or dependence, nicotine dependence (Patient Register)
• Traffic accidents and offences: accidents were those that involved death and/or emergency inpatient or outpatient hospital visits (Patient Register and Cause of Death Register). Offences were those that involved conviction or suspicion of crimes against the Road Traffic Offences Act (National Crime Register, Register of People Suspected of Offences)

Analysis

• Analyses involved comparing outcomes across both treatment and no-treatment populations, as well as using a within-subjects design which allowed each patient to act as their own control which minimises confound and selection effects.

Results

• During the study period (approx. 3 years, 1 month) relative rates for outcomes were:
  • Varenicline: 5.4% were suspected of a crime (vs. 4% no treatment)
  • Varenicline: 4.6% were diagnosed with a new psychiatric disorder (vs. 2.2%)
  • Varenicline: 0.9% received care for suicidal behaviour (vs. 0.3%)
  • Varenicline: 1.4% rate of serious traffic accidents (vs. 1.4%)
• Unadjusted comparisons between the treatment/no-treatment populations found that those on varenicline had significantly higher incidence of new psychiatric disorders, suicidal behaviour, suspected crime and crime conviction (but not traffic accidents or offences). When adjusting for age and sex, those on varenicline showed a higher incident rate on all outcome measures.
• When within subject analysis was conducted (i.e., individuals acted as their own control), varenicline was only associated with a higher rate of new psychiatric conditions, specifically anxiety and mood disorders (not psychosis). In addition, this hazard was much less in the within, compared with between, subject analysis (1.2 vs. 2.0).
• Further sensitivity analyses focused on comparing those with and without pre-existing psychiatric conditions, and found that the positive association between varenicline treatment and anxiety/mood disorders only occurred in those with pre-existing psychiatric conditions.
• To assess the potential confound of the nicotine withdrawal syndrome, comparisons were made between those who received varenicline or buproprion during follow-up. Those treated with varenicline had a reduced risk for mood disorders relative to those treated with buproprion and there was no difference in incidence of anxiety disorders.

Conclusions

The authors concluded that previous suggestions that varenicline could be associated with criminal and suicidal behaviour, traffic accidents and offences appear unfounded. It is possible that any associations are actually due to the fact that tobacco smoking has been positively correlated with suicidal behaviour, aggression and impulsivity, and traffic-related accidents.

A significant link did remain between varenicline and incidence of anxiety and mood disorders in those with pre-existing psychiatric disorders. However, this risk was similar to (or smaller) than the risk between bupropion treatment and anxiety and mood disorders. Therefore, it is possible that this increased risk is actually due to the experience of the nicotine withdrawal syndrome (or some other non-identified factor) and not the specific pharmacotherapy.

Limitations

• Firstly, the use of Registers to collect outcome data will only include those incidents serious enough to require emergency and hospital care or police detection. Therefore these rates are likely to be underestimates.
• Secondly, the treatment data is based on prescriptions being collected, and therefore does not take into account (non)adherence issues.
• Finally, given that Sweden has a lower smoking rate than the European average (14 vs. 23%) there may be some limitations to the generalisability of results.

However, overall this is a thorough, well-controlled, large sample study, investigating several important outcomes while controlling for various potential confounds.

Summary

This evidence does not support the previous associations made between varenicline use and increased suicidal and criminal behaviour, and traffic accidents, which have resulted from case reports and post-marketing surveys.

Although a significant increased risk of mood and anxiety disorders was associated with varenicline treatment, this only occurred in those with pre-existing psychiatric conditions. Additionally, when comparing treatments, varenicline was not associated with a greater risk of mood and anxiety disorders relative to buproprion. This suggests that other factors may be related to risk of psychiatric disorder, e.g., the nicotine withdrawal syndrome, or patients may change their psychiatric medication regime when commencing treatment for substance dependence.

Future research, using robust scientific designs, will need to focus on clarifying these issues and take detailed records of nicotine withdrawal symptomology, other medication regimes, treatment adherence rates etc. Based on future findings, current regulatory guidelines/warnings and industry prohibitions concerning varenicline should be revisited.

If you need help

If you need help and support now and you live in the UK or the Republic of Ireland, please call the Samaritans on 116 123.

If you live elsewhere, we recommend finding a local Crisis Centre on the IASP website.

We also highly recommend that you visit the Connecting with People: Staying Safe resource.

Links

Primary paper

Molero Y, Lichtenstein P, Zetterqvist J, Gumpert CH, Fazel S. Varenicline and risk of psychiatric conditions, suicidal behaviour, criminal offending, and transport accidents and offences: population based cohort study. BMJ 2015 Jun 2;350:h2388. doi: 10.1136/bmj.h2388.

Other references

Cahill K, Stead LF, Lancaster T. Nicotine receptor partial agonists for smoking cessation. Cochrane Database of Systematic Reviews 2012, Issue 4. Art. No.: CD006103. DOI: 10.1002/14651858.CD006103.pub6.