Searching for more effective smoking cessation treatment.

The study by Rose and Behm in this issue (1) is a remarkable first attempt to improve smoking cessation by selecting treatment for individual patients based on an initial test period. Smoking cessation treatments specify a quit date, usually 2 weeks after the initiation of treatment. Rose and Behm used this 2-week period to identify those patients who would likely respond to nicotine replacement therapy, the simplest and safest treatment. The test was to identify those patients who decreased their smoking by 50% during open-label treatment with a nicotine transdermal patch.Most patients receivedone 21-mgpatchdaily, but heavier smokers received two patches. The one-third of patients who decreased their cigarette consumption by 50% or more then continued on the patch for 12 weeks. Half these patients remained abstinent at 12 weeks, and 22% remained abstinent at 6months. The favorable prognosis for patients who immediately decreased smoking with nicotine patch treatment replicates earlier findings from Rose and colleagues (2). The unique feature of this study is the double-blind comparison of three interventions for patients who did not decrease their smoking by 50% during the prequit phase. One-third were continued on nicotine patch, one-third received the nicotine patch plus bupropion, and one-third were switched to varenicline. Bupropion was titrated to 150 mg twice daily and varenicline to 1mg twice daily. The treatment period lasted 12 weeks, and approximately 100 patients were in each group. At 12 weeks, the abstinence rate for treatment with nicotine patch alone was less than 7% for patients who had not decreased their smoking before the quit date, compared with 50% for those patients who had decreased smoking before the quit date. Patients who had bupropion added to the nicotine patch achieved a response rate of 19%, and patients whowere switched to varenicline achieved a response rate of 12%, the latter effect not significantly different from the nicotine patch condition. At 6 months, among the patients who had not responded during the prequit phase, 7% of those on nicotine patch alone, 17% of those who received bupropion in addition, and 16% of those on varenicline were abstinent. Thus, the addition of bupropion to the nicotine patch for patients who did not respond quickly to patch alone significantly increased the response rate. Varenicline had longer-term effects that were also significantly better than the patch alone for nonresponding patients. In a second phase of the study, a second intervention wasmade to try to “rescue” patients who relapsed during the first postquit week despite a favorable prequit response. This intervention was not as successful, unfortunately. However, the small number of participants in each group, approximately 30, may have contributed to the inconclusive findings. Clinical lessons from the study include the finding that patients’ self-report of cigarette consumption closely correlated with the results of carbon monoxide monitoring and that percent decrease in smoking predicted the clinical course better than absolute levels of cigarette consumption. The study design is complicated