Smoking cessation trials/Six of the authors respond

• Published in: Canadian Medical Association journal (CMAJ) 2008-11, Vol.179 (10), p.1037
• Main Authors: Polito, John R, Filion, Kristian B, Bélisle, Patrick, Joseph, Lawrence, Paradis, Gilles, Rinfret, Stéphane, Eisenberg, Mark J
• Format: Article
• Language: English
• Subjects: Systematic review
• ISSN: 0820-3946; 1488-2329

I submit that the meta-analysis by [Mark J. Eisenberg MD MPH] and colleagues on pharmacotherapies for smoking cessation1 is grounded in a false premise, namely that researchers were somehow able to hide the onset of nicotine withdrawal symptoms from control group members, whose previous quitting history had taught them exactly how withdrawal felt (a rising tide of anxieties, anger, dysphoria, concentration difficulty and sleep fragmentation within 24 hours of quitting), and that researchers found a way to mask the reduction of withdrawal syndrome for intervention group members. Mooney and colleagues found that studies of nicotine replacement therapies are generally not blind in that participants correctly guess assignment at rates significantly above chance.2 When this finding is combined with the meta-analytic finding by Eisenberg and colleagues that smoking cessation with pharmacologic treatment is nearly always more successful than cessation without pharmacologic treatment in clinical trials and the fact that cessation with pharmacological treatment has failed to be more successful than cessation without such treatment in nearly all of real-world surveys conducted to date,3 it strongly suggests that the pharmacologic treatment of chemical dependency may be the only known research area in which blinding is impossible. We thank [John R. Polito JD] for his discussion of the importance of blinding in clinical trials. Blinding is undoubtedly a key component of the validity of inferences drawn from randomized controlled trials. For this reason, it is included in tools to assess the quality of clinical trials, such as the Jadad scale.1 In our meta-analysis of randomized controlled trials of smoking cessation pharmacotherapies, we restricted our study to double-blind trials so that only trials of the highest quality would be included.2 We agree with Polito that maintaining blinding may be difficult, particularly if study participants experience withdrawal symptoms. However, although the importance of blinding is well established, assessing the integrity of blinding remains more controversial.