FROM THE LANCET
Neuropsychiatric adverse events do not increase significantly in smokers treated with either varenicline or bupropion, a large cohort study shows.
Both bupropion and varenicline have been tied to long-term smoking cessation in observational studies and randomized trials. However, concerns about adverse neuropsychiatric events, including aggression and suicidality, have been raised. Furthermore, data are limited on the safety of the medications in smokers with known psychiatric conditions.
At the request of the Food and Drug Administration, Dr. Robert M. Anthenelli and his colleagues conducted a randomized, double-blind, triple-dummy, placebo- and active-controlled trial to assess bupropion and varenicline in motivated smokers with and without psychiatric diagnoses for 12 weeks. The efficacy endpoint in the multinational trial, called the Evaluating Adverse Events in a Global Smoking Cessation Study (EAGLES), was abstinence for 9-12 weeks. The primary endpoint was adverse neuropsychiatric events, reported Dr. Anthenelli of the psychiatry department at the University of California, San Diego.
In total, 8,144 participants were randomized to either a nonpsychiatric (n = 4,028) or a psychiatric (n = 4,116) cohort. Men made up 44% of the study population, and the average age was 46.5 years. Most participants were white (82%) and American (52%). The psychiatric cohort included participants with diagnoses of primary mood disorders, anxiety and psychotic disorders, and borderline personality disorders, and 49% reported treatment with a psychotropic medication (Lancet. 2016 Apr 22. doi: 10.1016/S0140-6736 ).
Overall, the incidence of neuropsychiatric adverse events was similar in the bupropion (4.5%), varenicline (4.0%), nicotine patch (3.9%), and the placebo (3.7%) groups. However, more neuropsychiatric events were reported in the psychiatric cohort than the nonpsychiatric cohort (5.8% versus 2.1%, P less than .0001). Likewise, the psychiatric cohort reported moderate and severe neuropsychiatric adverse events more often in the bupropion group (6.7% versus 2.2%), varenicline (6.5% versus 1.3%), nicotine patch (5.2% versus 2.5%), and placebo groups (4.9% versus 2.4%) than the nonpsychiatric cohort.
In the nonpsychiatric cohort, the risk differences for moderate and severe neuropsychiatric adverse events were –1.28 (95% confidence interval, –2.40 to –0.15) for varenicline vs. placebo and –0.08 (95% CI, –1.37 to 1.21) for bupropion vs. placebo. In the psychiatric cohort, the risk differences for moderate and severe neuropsychiatric adverse events were 1.59 (95% CI, -0.42 to 3.59) for varenicline-placebo and 1.78 (95% CI, -0.24 to 3.81) for bupropion-placebo.
Rates of abstinence were higher in the participants who received varenicline, compared with placebo (OR, 3.61; 95% CI, 3.07-4.24), bupropion (OR, 1.75; 95% CI, 1.52-2.01), and the nicotine patch (OR, 1.68; 95% CI, 1.46-1.93).
The most common adverse events reported included abnormal dreams, headache, insomnia, and nausea.
Dr. Anthenelli and his associates noted several limitations. For example, participants in the psychiatric cohort were stable or in remission; they were restricted to particular psychiatric diagnoses; and participants with current substance abuse or risk for suicide were excluded.
However, they said the EAGLES trial results provide “further evidence that varenicline and bupropion can be used safely by psychiatrically stable smokers,” they wrote. “Although varenicline appears to be the most effective single pharmacotherapy available, all of the first-line medications – varenicline, bupropion, and nicotine patch – are efficacious, compared with placebo.”
The authors report relationships with several pharmaceutical companies, including Pfizer and GlaxoSmithKline. The study was funded by Pfizer and GlaxoSmithKline.
