Interventions for smoking cessation: the outcomes are determined by the questions asked.

There is an increasing awareness of the health risks associated with smoking among the general public. However, the rates of smoking in Western countries have been remarkably resistant to change, and in developing countries the rates are increasing. Many interventions of the health promotion variety are being explored, while the tobacco industry is just as interested in developing more effective techniques to promote smoking. There have been vigorous attempts to find the optimum ways to help people quit. There is a broad awareness among professionals that interventions to help people who want to stop smoking benefit from both behavioural and pharmacological components. In the current decade, most of the advances have been pharmacological. A decade ago, interventions that were primarily behavioural were predominant. For example, a meta-analysis by VISWESVARAN and SCHMIDT [1], included the results of 633 studies conducted before nicotine replacement was commonly available. Nicotine replacement (nicotine gum) was used in 6% of these studies. The most effective interventions were multi-component programmes (35% success rate), hypnosis (36%), smoke aversion techniques (31%), group withdrawal clinics (30%), and acupuncture (30%). The least effective were self-care (15%), nicotine gum (16%), and physician advice (7%). Cessation results for controls averaged 6%. Subsequent meta-analyses have more often focused on the use of nicotine replacement therapies. CEPEDA-BENITO [2], in a meta-analysis of 33 studies of the efficacy of nicotine gum, found that the only studies where the gum affected success in the long-term were those where it had been combined with intensive behavioural treatments. SILAGY et al. [3] considered the results of 53 trials involving nicotine replacement therapy and concluded that it was effective in different forms, but largely independent of the intensity of additional support provided. FIORE et al. [4] included 17 nicotine patch efficacy studies that were double-blind, placebo-controlled with random assignment of subjects. They found that active patch subjects were more than twice as likely to quit smoking as were participants using a placebo patch. This effect was present at both high and low intensities of counselling. Intensive behavioural counselling had a modest positive impact on quit rates, which was stronger at the end of treatment than after six months. In the metaanalysis conducted by LI WAN PO [5], the extent of behavioural support in each study is described, but it is not included as a variable in the meta-analysis. It is easy to conclude from the more recent of these reports that the behavioural counselling component of these programmes is largely ineffectual. However, nicotine patch studies are often funded by the industry that manufactures the patch, which has a relatively small interest in demonstrating the contribution of behavioural interventions. Furthermore, investigators often appear to emulate the level of counselling that would be delivered in a physician's office, and it is unlikely that intensive counselling would be offered in such a setting. Studies have compared various nicotine delivery systems, various concentrations of nicotine, and the combination of nicotine with several other agents. Such studies are both time consuming and expensive to conduct, but are necessary in order to gauge the progress made by current revisions to "best practice" in the area of interventions for smoking. The results reported by TùNNESEN et al. [6] in this issue of the European Respiratory Journal add to this literature. In the TùNNESEN et al. [6] Collaborative European AntiSmoking Evaluation (CEASE) trial, the investigators at 36 clinical centres collaborated to provide results from 3,575 adults who were attempting to stop smoking. Participants were given 25 or 15 mg or placebo patches, for 8 or 22 weeks in a randomized double-blind design. The rate of sustained smoking cessation was followed for 1 yr. The investigators found that dose was related to cessation success, but duration of treatment was not. In the CEASE trial, there was a 10-fold difference between cessation rates in the active and placebo groups in the best of their clinics as compared to their least successful clinic. This may either be due to clinic differences at baseline or to different effects of the intervention in the clinics. It would be very interesting to rank the intensity of behavioural intervention in these clinics, and then assess whether clinics divided into quintiles of intensity of the behavioural programmes would find that intensity was reflected in their mean success rates. TùNNESEN et al. [6] report 12-month sustained cessation rates of around 16% with the 25 mg patch, 13% with the 15 mg patch, and 10% with placebo. The Lung Health Study, for example, reported 12-month sustained cessation rates, biochemically verified, of approximately 35% with ad libitum use of 2 mg nicotine gum and intensive behavioural intervention, and 10% for usual care (controls) [7]. It is not possible to attribute the differences in cessation rates in the *Dept of Community Health Sciences and Faculty of Medicine, University of Manitoba, Winnipeg, Manitoba, Canada