The Lancet April 22, 2016

Neuropsychiatric safety and efficacy of varenicline, bupropion, and nicotine patch in smokers with and without psychiatric disorders (EAGLES): a double-blind, randomised, placebo-controlled clinical trial

Anthenelli RM, Benowitz NL, West R, et al.

Bottom Line

The EAGLES trial demonstrated that varenicline and bupropion do not significantly increase the risk of serious neuropsychiatric adverse events compared to nicotine patch or placebo, and confirmed varenicline as the most effective pharmacotherapy for smoking cessation.

Key Findings

1. In the psychiatric cohort, the incidence of the primary composite neuropsychiatric safety endpoint was comparable across all treatment arms: varenicline (6.5%), bupropion (6.7%), nicotine patch (5.2%), and placebo (4.9%).
2. In the non-psychiatric cohort, the safety endpoint incidence was low and similar across groups: varenicline (1.3%), bupropion (2.2%), nicotine patch (2.5%), and placebo (2.4%).
3. Risk differences for the primary safety endpoint versus placebo were not statistically significant (varenicline RD 1.59%, 95% CI -0.42 to 3.59 in the psychiatric cohort; RD -1.28%, 95% CI -2.40 to -0.15 in the non-psychiatric cohort).
4. For efficacy (continuous abstinence during weeks 9-12), varenicline was significantly superior to placebo (OR 3.61), nicotine patch (OR 1.68), and bupropion (OR 1.75).
5. Bupropion (OR 2.07) and nicotine patch (OR 2.15) were also significantly more effective than placebo in achieving continuous abstinence.
6. Across all cohorts, the most frequent adverse event associated with varenicline was nausea (25%), while insomnia was most common with bupropion (12%).

Study Design

Design
Double-Blind RCT
Double-Blind
Sample
8,144
Patients
Duration
24 wk
Median
Setting
Global, multicenter
Population Adult daily smokers (10 or more cigarettes/day) aged 18 to 75 years, partitioned into cohorts with and without a history of clinically stable psychiatric disorders (mood, anxiety, or psychotic disorders).
Intervention Varenicline (1 mg twice daily) or Bupropion SR (150 mg twice daily) administered for 12 weeks.
Comparator Nicotine patch (21 mg/day with a stepwise taper) or Placebo administered for 12 weeks.
Outcome A composite of moderate-to-severe neuropsychiatric adverse events (primary safety) and carbon-monoxide-confirmed continuous smoking abstinence during weeks 9 through 12 (primary efficacy).

Study Limitations

The trial excluded individuals with active, unstable, or untreated psychiatric illnesses, as well as those with a recent history of substance use disorders.
Patients with suicidal behaviors within the past year or a lifetime history of suicidal ideation with intent were excluded, limiting generalizability to the most high-risk psychiatric populations.
While massive in scale, the trial may still have been underpowered to detect extremely rare, catastrophic psychiatric events such as completed suicides (though the observed event rates were exceptionally low).

Clinical Significance

The EAGLES trial provided definitive, high-quality evidence that first-line smoking cessation medications (varenicline and bupropion) are neuro-psychiatrically safe for patients with stable mental health conditions. By directly refuting earlier fears of drug-induced suicidality, the findings removed a major barrier to prescribing life-saving cessation therapies in psychiatric patients—a population that suffers disproportionately from tobacco-related morbidity and mortality. This led directly to the FDA removing the 'black box' warnings from both varenicline and bupropion.

Historical Context

Following early post-marketing case reports linking varenicline and bupropion to severe psychiatric adverse events—including depression, aggression, and suicidality—the FDA placed stringent 'black box' warnings on both medications in 2009. These warnings severely curtailed prescribing practices, particularly for patients with underlying psychiatric illnesses who paradoxically have the highest rates of smoking and the most urgent need for effective cessation aids. To definitively address this, the FDA and the European Medicines Agency (EMA) mandated an unprecedentedly large, rigorous safety trial. The resulting EAGLES trial decisively demonstrated the neuropsychiatric safety of these drugs, prompting the FDA to remove the black box warnings in December 2016.

Guided Discussion

High-yield insights from every perspective

Med Student
Medical Student

Varenicline and bupropion are both effective for smoking cessation but operate via different mechanisms. How does varenicline's mechanism as a partial agonist at the alpha-4 beta-2 nicotinic acetylcholine receptor explain both its efficacy in reducing cravings and its ability to block the rewarding effects of smoking?

Key Response

Understanding partial agonism is a foundational pharmacology concept. Varenicline stimulates the receptor enough to prevent withdrawal and cravings (its partial agonist effect) but simultaneously blocks the binding of nicotine from a cigarette, preventing the massive dopamine spike that reinforces the smoking habit (its antagonist effect).

Resident
Resident

A 45-year-old patient with well-controlled major depressive disorder wants to quit smoking but is terrified to try varenicline because they read it causes suicidal thoughts. Based on the EAGLES trial, how should you counsel this patient and integrate this evidence into your shared decision-making regarding first-line therapy?

Key Response

Residents must translate evidence to patient counseling. The EAGLES trial explicitly demonstrated that varenicline does not increase the incidence of clinically significant neuropsychiatric adverse events compared to placebo or nicotine patch, even in patients with stable psychiatric conditions, allowing clinicians to confidently recommend it as a safe and highly efficacious first-line therapy.

Fellow
Fellow

While the EAGLES trial showed no significant difference in neuropsychiatric safety between treatments within the psychiatric cohort, the baseline rate of these events was inherently higher in this cohort compared to the non-psychiatric group. How should this nuanced finding influence the frequency and structure of follow-up for a patient with severe bipolar disorder initiated on varenicline?

Key Response

Fellows need to recognize that 'no difference vs placebo' does not mean 'zero risk.' Patients with underlying psychiatric illness have a higher baseline risk of psychiatric exacerbations (e.g., from nicotine withdrawal itself) and still require vigilant, structured psychiatric monitoring during cessation attempts regardless of the pharmacotherapy chosen.

Attending
Attending

The FDA originally placed a black box warning on varenicline based on post-marketing surveillance, which the EAGLES trial ultimately reversed. What does this reversal teach us about the limitations of spontaneous adverse event reporting systems compared to large-scale, actively controlled randomized trials when evaluating drugs for populations with high background rates of pathology?

Key Response

Attendings must contextualize evidence history. Post-marketing data is vulnerable to confounding by indication and background noise (nicotine withdrawal inherently causes mood changes, and psychiatric patients inherently have mood fluctuations). A placebo-controlled trial was absolutely required to separate the drug effect from withdrawal symptoms and baseline disease progression.

Scholarly Review

Critical appraisal through the lens of expert reviewers and guideline development

PhD
PhD

The EAGLES trial utilized a complex, customized composite endpoint for neuropsychiatric adverse events designed in collaboration with the FDA. What are the methodological strengths and vulnerabilities of constructing a novel composite safety endpoint rather than relying on standard MedDRA hierarchies, and how might this affect the trial's power and type-II error rates for rare events?

Key Response

A broad composite endpoint captures a wide safety net, increasing statistical power to detect any event. However, if the composite is driven by lower-severity, high-frequency events like mild agitation, it could dilute the signal of rare, severe events like completed suicide, potentially risking a Type II error for the most critical individual safety outcomes.

Journal Editor
Journal Editor

The psychiatric cohort in the EAGLES trial required patients to be clinically stable to participate. As a peer reviewer, how does this inclusion criterion threaten the external validity of the study, and does the blanket claim that varenicline is safe in 'patients with psychiatric disorders' overreach the actual trial population?

Key Response

Editors must look for overgeneralization. The trial excluded patients with active, unstable psychiatric disease or recent substance abuse. Therefore, the safety data strictly applies to stable psychiatric patients, and extrapolating these findings to an acutely psychotic or actively suicidal patient is a significant leap that authors must carefully caveat.

Guideline Committee
Guideline Committee

Given the definitive safety and superior efficacy data from the EAGLES trial, how should current clinical practice guidelines formally upgrade the recommendation for varenicline in patients with stable psychiatric comorbidities, and should insurance-driven step-therapy (requiring NRT failure first) be officially opposed by professional societies?

Key Response

Guideline committees evaluate whether new evidence mandates a shift in first-line status. The EAGLES trial provides Level 1a evidence that varenicline is safe and superior to NRT, strongly supporting guidelines (like those from the ATS or ACC) to explicitly recommend varenicline as the uncontested first-line agent, effectively abolishing the clinical justification for NRT step-therapy.

Clinical Landscape

Noteworthy Related Trials

2006

Varenicline vs. Bupropion Trial

n = 1027 · JAMA

Tested

Varenicline 1mg twice daily

Population

Healthy adult smokers

Comparator

Bupropion SR or placebo

Endpoint

Continuous abstinence at weeks 9-12

Key result: Varenicline produced significantly higher smoking cessation rates compared with both bupropion SR and placebo.
2014

Maintenance Varenicline Trial in Schizophrenia

n = 87 · JAMA

Tested

Varenicline maintenance therapy

Population

Smokers with schizophrenia or bipolar disorder

Comparator

Placebo

Endpoint

7-day point prevalence abstinence at 52 weeks

Key result: Maintenance varenicline therapy improved long-term smoking abstinence without worsening psychiatric symptoms.
2015

EVITA Trial

n = 302 · Circulation

Tested

Varenicline

Population

Smokers hospitalized with acute coronary syndrome

Comparator

Placebo

Endpoint

Continuous smoking abstinence at 24 weeks

Key result: Varenicline significantly improved smoking cessation rates in post-ACS patients without increasing cardiovascular adverse events.

Tailored to your role

Want this tailored to you?

Add your specialty or training stage to get role-specific takeaways and more questions.

Personalize this analysis