Nix Box Warning on Smoking Cessation Drug, FDA Panel Says

Pauline Anderson

September 16, 2016

A joint US Food and Drug Administration (FDA) advisory committee has voted to remove a box warning about possible neuropsychiatric side effects associated with use of a prescription smoking cessation medication.

Members of the Psychopharmacologic Drug Advisory Committee and the Drug Safety and Risk Management Advisory Committee concluded that in light of results of a large postmarketing randomized clinical trial, the box warning on varenicline (Chantix. Pfizer) is not warranted, despite concerns about the study raised by the FDA in a report.

However, committee members stressed that the precedent-setting removal of the box warning should not send the message that this drug is now completely safe.

Varenicline is a partial α4β2 acetylcholine nicotine receptor agonist that was approved in May 2006 as an aid to smoking cessation. Another drug, buproprion hydrochloride, an aminoketone antidepressant originally approved under the name Wellbutrin (GlaxoSmithKline), was approved for smoking cessation under the name brand name Zyban in 1997.

On the basis of postmarketing adverse event reports, Chantix and Zyban carry box warnings on the risk for serious neuropsychiatric events, such as suicidal thoughts and suicide.

In 2008, the FDA required Pfizer and GlaxoSmithKline to conduct a placebo-controlled postmarketing safety trial to further characterize the risk for neuropsychiatric adverse events and to evaluate whether a prior history of psychiatric illness was a risk modifier. Patients with a history of psychiatric illness did not participate in the initial clinical efficacy trials.

The study, a double-blind, active- and placebo-controlled parallel group trial, was designed to assess the safety and efficacy of varenicline 1 mg BID and bupropion hydrochloride 150 mg BID for smoking cessation. The primary comparisons were varenicline vs placebo and bupropion vs placebo. Nicotine replacement therapy (NRT) was included as an active control, and study medications were given via a triple-dummy design.

The treatment continued for 12 weeks and was followed by a nontreatment follow-up phase of another 12 weeks. About 2000 patients in each of four treatment arms were included in the trial, for a total of just over 8000 patients.

Reliable Data?

Researchers classified participants into one of two cohorts ― those with an established and stable diagnosis of psychiatric disorder, and those without a diagnosis of a psychiatric disorder.

The primary outcome was a novel composite neuropsychiatric symptom (NPS) adverse event (AE) endpoint that had 16 components, including agitation, aggression, panic, mania, depression, psychosis, depression, and anxiety, among others.

Information on these symptoms was collected through a variety of means, including a Neuropsychiatric Adverse Event Interview (NAEI), which is a semistructured interview administered by trained personnel.

According to study sponsor Pfizer, results showed that for the overall study population, the percentages of NPS AEs were as follows: 4.0% for varenicline; 4.5% for bupropion; 3.9% for NRT; and 3.7% for placebo. In the nonpsychiatric group, the percentages were as follows: 1.3% for varenicline; 2.2% for bupropion; 2.5% for NRT; and 2.4% for placebo. For the psychiatric group, the percentages were as follows: 6.5% for varenicline; 6.7% for bupropion; 5.3% for NRT; and 4.9% for placebo.

In the nonpsychiatric group, serious or significant NPS AEs occurred in all treatment groups, but the incidences were similar across all treatment arms. In the psychiatric group, such AEs also occurred in all treatment groups but were more frequent in the varenicline and bupropion groups.

The vast majority of events had an impact on patient functioning, although they were not of a serious nature and were usually transient.

During the advisory committee meeting, Pfizer representatives stressed that varenicline is the most efficacious smoking cessation aid available and that the box warning should be removed, as it may deter appropriate use of this drug.

Several speakers stressed that smoking is an addiction, that more people with mental illness smoke, and that quitting smoking, in and of itself, is not without risk.

The FDA raised several concerns about data collection during the study. For example, Celia Winchell, MD, clinical team leader, Division of Anesthesia, Analgesia, and Addiction Products, Office of Drug Evaluation, Office of New Drugs, Center for Drug Evaluation and Research, FDA, noted that the NAEI was used as a "checklist," not as a tool to guide a semistructured interview, as it was intended. In addition, she said that a full report of patient experiences was not reliably captured, and it was not always clear what patients actually said during interviews.

There were also problems with data coding, with the assessment of severity being "very inconsistent," said Dr Winchell. Narratives of events that were similar were at times described as mild and at other times as severe, she said, adding that some information was not collected or not reported.

Dr Winchell outlined other problems related to data quality. For example, in a number of sites, unqualified people performed procedures that were supposed to be performed by mental health professionals. There was some question about the financial disclosures of investigators at some sites, and two sites among 26 audited by the sponsor were found to be unreliable.

An FDA sensitivity analysis also showed unexpected variability, or heterogeneity, in the primary event rate across sites that was difficult to explain.

However, in the end, the FDA essentially came to a conclusion similar to that of the trial's sponsor about the efficacy and safety of varenicline.

During discussions, committee members generally agreed that the design of the study was fairly good, although some raised concerns about quality control across multiple sites with multiple investigators and about cultural and linguistic variations.

Some highlighted the fact that the population with the most severe psychiatric illnesses was underrepresented and that the study lacked power to look at rare events, such as suicide. There was also some concern that 15% to 20% of enrolled patients had in the past been exposed to the drugs being investigated.

Many members were worried about the validity of outcome measures, particularly the NAEI, which was designed to have "narrative" as a complementary component.

They also discussed "noise" in the data relating to the relatively high level of agitation in study participants. Such agitation could have been caused by exposure to the drug or by nicotine withdrawal, which makes it difficult to interpret this part of the primary endpoint.

The Vote

Stephen Marder, MD, professor and director, Section on Psychosis, Semel Institute for Neuroscience, University of California, Los Angeles, said he would have liked to have seen outcome data related to different psychiatric diagnoses.

"Most psychiatrists don't see psychiatric patients as a single group; they see them as people with individual illnesses. It would be useful in the future to get data on risks related to what illness a patient has," he said.

The FDA also provided an update on observational studies. Committee members heard that overall in these studies, most outcomes were not significant, fatal and nonfatal self-harm events were rare, and there was no evidence of an increased risk for most severe NPS events in patients receiving varenicline.

Ten of 19 committee members voted in favor of eliminating the box warning, four voted to modify the warning, and five voted to keep the warning.

Several members stressed the need to carefully watch patients with a psychiatric condition who take a prescription smoking cessation drug and that labeling should reflect the continued need for vigilance.

In explaining why he voted to remove the box warning, Dr Marder noted, "Over the years, with additional data that's come up, there hasn't been anything that sort of reinforced the initial anecdotal spontaneous reports." That, he said, "made me think that the signal is just not strong enough to justify a black box."

But Dr Marter did wonder why the rates of NPS AEs were so low in patients receiving placebo when people quitting smoking "are supposed to feel crappy."

Tobias Gerhard, PhD, associate professor, Department of Pharmacy Practice and Administration, Rutgers University, New Brunswick, New Jersey, also voted to remove the box warning. The study, he said, "allows us to quantify the risk compared to what was known at the time the box was put in place," and that "makes it pretty clear that in this specific drug and indication, the benefits outweigh the risks pretty clearly."

Sean Hennessy, PharmD, PhD, professor of epidemiology in biostatistics and epidemiology, University of Pennsylvania, in Philadelphia, emphasized that because the study was underpowered for NPS AEs, "the only evidence associated with serious neuropsychiatric events are case reports, some of which are very convincing, but there are also very convincing case reports in people with placebo, so I think that reduces the convincingness of the spontaneous reports."

Dr Hennessy, who also voted for removal of the box warning, added, "If varenicline does cause serious neuropsychiatric events, it does so in a relatively small proportion of people, while smoking causes severe adverse events in a very high proportion of people."

Elaine H. Morrato, DrPH, associate dean for public health practice and associate professor, Department of Health Systems, Management and Policy, Colorado School of Public Health, in Aurora, was among five members who voted to keep the box warning. Because of the "precedent-setting nature" of removing it, she said she would "feel more comfortable" if more time was devoted to reviewing all the analyses.

As with many of her committee colleagues, Dr Morrato said she worries about the "unintended consequence" that removing this warning may have.

"It would send the message that it's safe, and in this case, it requires careful messaging that a box warning being removed doesn't mean that," she said.

But Dr Hennessy warned of the implications of keeping the box warning.

"People are concerned about the message that taking away the box warning would have, but continuing the black box warning would also send a message that there's a continuing serious safety problem with the drug. And that would continue to promote underuse of an effective smoking cessation therapy, particularly in those who most need it ― those with serious mental health conditions."

According to the FDA, it is highly unusual that it would remove a box warning and that it has only done so in isolated cases.


Comments on Medscape are moderated and should be professional in tone and on topic. You must declare any conflicts of interest related to your comments and responses. Please see our Commenting Guide for further information. We reserve the right to remove posts at our sole discretion.
Post as: