US drug regulator the Food and Drug Administration (FDA) has issued warnings about possible side effects of taking varenicline. Varenicline (marketed as Chantix in the US, and Champix in the UK) is a drug which can help people stop smoking. The FDA’s warnings are out of date, do not reflect the scientific evidence, and urgently need to be updated. They may have led the public and clinicians to incorrectly believe that there was strong scientific evidence that varenicline causes or exacerbates mental health problems. Given varenicline is the most effective way to successfully stop smoking, these safety concerns may have resulted in fewer people successfully stopping smoking.
A brief history FDA’s warnings over varenicline
In May 2006 the FDA approved varenicline for use as a medication to help people stop smoking. This was the first treatment licensed for smoking cessation since 1997, and initial evidence suggested that people who took it were more likely to successfully stop smoking. However, it was not long before concerns were raised that people prescribed varenicline were experiencing mental health side effects. Newspapers reported a series of anecdotes that people taking varenicline had suicidal behaviour. The FDA’s Adverse Event Reporting System (AERS) suggested patients prescribed varenicline were more likely to report suicidal behaviour than patients prescribed other medications. The key question was, are these differences a result of taking varenicline, or is it just that patients prescribed it had a higher risk of these events even before they were prescribed treatment? Were they due to correlation or causation?
The FDA attempted to answer this question in a safety review that was published in February 2008. They concluded that “[varenicline] may cause worsening of a current psychiatric illness even if it is currently under control and may cause an old psychiatric illness to reoccur.” In July 2009, after pressure from the media and consumer groups, the FDA went further, mandating that varenicline carry a “Black Box warning”. Black Box warnings are the FDA’s strongest warning about possible side effects of drugs. The Black Box warning for varenicline highlights “the risk of serious neuropsychiatric symptoms in patients using these products”. Symptoms included changes in behaviour, hostility, agitation, depressed mood, suicidal thoughts and behaviour, and attempted suicide.
The FDA then commissioned two large observational cohort studies to examine the risks associated with varenicline. They also instructed Pfizer to conduct a large randomised trial to investigate the side effects of varenicline in people with mental health problems.
Results from the observational studies were reported in October 2011. In both studies people prescribed varenicline had similar risks of being hospitalised for psychiatric problems as patients prescribed nicotine replacement therapy – in other words, there was no evidence of increased risk. However, these were observational studies, rather than experimental (randomised trials), so could not prove that varenicline did not cause adverse outcomes. However, most epidemiologists would regard these studies as more reliable evidence than the anecdotal case reports and newspaper reports which originally raised concerns about these risks. Nevertheless, in response to this new evidence, the FDA “determined that the current warnings in the varenicline drug label, based on post marketing surveillance reports, remain appropriate” and did not remove the Black Box warning.
Meanwhile, further observational studies were amassing evidence about whether varenicline use is associated with adverse events. In 2009, our colleagues found no evidence that patients in UK prescribed varenicline had higher risks of suicidal behaviour. In 2013, we updated this evidence to include over 100,000 patients, and again found that patients prescribed varenicline had similar risks of suicidal behaviour and similar rates of self-harm. In 2015, we systematically reviewed the literature of randomised controlled trials, and again found that participants given varenicline had similar risks of suicide and self-harm as those given placebo.
In March 2015 the FDA issued a further update. Despite mounting evidence about the safety of varenicline, they did not change their warnings. In 2016, the randomised controlled trial the FDA commissioned reported its findings. This trial randomised 8,144 people to receive varenicline, nicotine patch, bupropion or placebo, and found that people allocated to varenicline rather than nicotine replacement products experienced fewer moderate and severe mental health adverse events. The European Medicine Agency, which regulates pharmaceuticals in Europe, recently updated the information it provides about varenicline. In contrast, to date the FDA has refused to update its warnings.
Correlation does not equal causation
Patients who take medications are different from the general population. This means that even if patients prescribed a drug are more likely to have a particular adverse event this does not necessarily prove that these events are caused by the drug. In the case of varenicline, people who smoke are different from the general population on average; they are usually poorer, sicker, and more likely to have mental health problems. So the fact that patients taking varenicline reported higher rates of suicidal behaviour than patients prescribed other drugs in the FDA Adverse Events Reporting System, may have been because these patients were at higher risk of these events, before taking varenicline. This means these adverse events are unreliable evidence of the difference when taking varenicline or other nicotine replacement products.
Spurious safety warnings can be extremely harmful
In 1998 a paper was published which suggested that there might be an association between taking the measles, mumps and rubella (MMR) vaccine and developing autism. These claims were false, and multiple, large, well-conducted studies have proven that these claims were false, and the MMR vaccine is not associated with, and does not cause autism. Nevertheless, these warnings led to a substantial fall in vaccination rates, and an increase in cases of measles. So it is vital for regulators to distinguish between side effects which are caused by drugs, and potential events which the drugs do not cause. These could be events that are correlated with a medication, or risks for which the regulator does not have any reliable evidence.
In 2007 3.9 million patients took varenicline in the US. In 2012, after the safety warnings were issued, only 1.2 million patients took varenicline. Thus a potential cost of issuing safety warnings is that fewer patients take otherwise effective medications.
What should we do now?
From the first warnings about potential adverse effects of varenicline on mental health in November 2007 to today, 4.1 million people are likely to have died from smoking related disease in the US alone. The FDA’s Black Box warning of mental health problems associated with varenicline is inconsistent with the latest scientific evidence. Ultimately, it could discourage people from using the most effective smoking cessation treatment. In 14th of September 2016, the FDA’s expert advisory group for drug safety met and again reviewed the evidence about the side effects associated with varenicline. The vote was close: they voted 10 to 9 in favour of advising the FDA to remove its warning. However, given the committee’s lack of consensus, it is not clear if the FDA will follow their guidance. In our opinion, given the evidence, the FDA needs to urgently revise this safety warning.
Dr Kyla Thomas is a National Institute for Health Research funded Clinical Lecturer in Public Health Medicine based at the University of Bristol and an Honorary Consultant in Public Health. She tweets sporadically @drkyla_thomas. Dr Neil Davies is an epidemiologist based in the MRC Integrative Epidemiology Unit at the University of Bristol. He occasionally tweets from @nm_davies. These views are their own.
A longer more detailed version of this blog post was published in the Addiction journal available here.
By failing to update their warnings, drug regulators are letting patients down
