Statistics from Altmetric.com
In this issue of Thorax Aubin and colleagues1 report a further trial from the Varenicline Phase III Programme (see page 717). The trials supporting registration contrasted bupropion with varenicline in a double placebo design.2 3 This study examines the efficacy of varenicline against nicotine replacement therapy (NRT). In many countries, including the UK, bupropion is rarely used and NRT is the predominant treatment offered in general practices and in specialist smoking cessation clinics. It is not practical to obtain placebo NRT, so this trial was of an open-label design. This publication follows a study by Stapleton et al4 with historical controls which showed that varenicline is superior to NRT in achieving abstinence and in reducing withdrawal phenomena such as urges to smoke and withdrawal symptoms.
Varenicline is licensed for smoking cessation around the world, but in the UK the National Institute for Health and Clinical Excellence (NICE) makes decisions about whether the benefits from licensed medications are worth the costs and therefore whether such interventions can be used in the National Health Service (NHS). NICE approved varenicline in 2007,5 but many local primary care organisations have limited the use of varenicline by making it a second-line treatment choice. The main reason given is the lack of trial data against the key competitor, NRT, but many have felt that the main reason was actually cost. Varenicline costs slightly more per treatment course than NRT or bupropion and considerably more than nortriptyline, an effective6 but unlicensed medication. The trial by Aubin and colleagues showing that varenicline is more effective than NRT and the trials of varenicline versus bupropion mean that varenicline must be considered a first-line choice in smoking cessation treatment. The recently published NICE guidance on smoking cessation affirmed this view,7 and it should be unnecessary for patients to have to try and to have failed with other medications before being offered varenicline in the NHS or in other health systems.
Although the statistically non-significant difference in continuous abstinence at 1 year in this trial can be considered as lack of evidence of long-term superiority, this would be clutching at straws. If we put together the data showing that varenicline is clearly more effective than bupropion8 with data that bupropion and NRT are roughly equally effective,9 the data from this trial1 and from the other published head-to-head comparison with NRT4 and data from the UK stop smoking services, we see a consistent picture. Varenicline is clearly more effective. Furthermore, although it is somewhat more expensive per treatment course, the difference in cost is small and it is likely to be the most cost-effective medication available for smoking cessation.4 Thus, we are in the happy position of having a licensed medication approved for use by NICE that is more effective and more cost-effective than competitors.10 Should varenicline become the only medication to be used in smoking cessation?
In most areas of medicine the most effective and cost-effective medication would mean other choices were rarely used, but this is unlikely to be the case in smoking cessation. We need to remind ourselves that smoking is a behaviour largely explained by addiction, and one of the hallmarks of addiction is reinstatement after abstinence. Thus, in the study by Aubin et al1 which probably took place in the best clinics, only a quarter of smokers remained smoke-free for 1 year. One-third of these will probably resume smoking again in the next 2 years.11 Thus, in the best hands, 8 or 9 out of 10 treatment episodes end in failure. According to NICE,12 smoking cessation treatment remains “among the most cost-effective of all healthcare interventions” because treatment is cheap and because the benefits of stopping smoking are so large in comparison to continuing. However, many failures occur early in treatment—nearly half those on varenicline had resumed smoking by the end of treatment in the trial by Aubin et al.1 In smoking cessation practice, then, we see a cadre of patients who have tried to stop many times and have, in effect, performed unblinded n-of-1 trials of all available smoking cessation medications. These patients are experts in which treatment works for them. Unlike many pharmacotherapies, patients can feel treatments for tobacco addiction working. Patients need to feel confident and content in their medication choice. In most conditions the main effort patients expend in treatment is remembering to take their drugs but, in smoking cessation, considerably more effort than this is usually required to achieve success. Confidence in the medication is probably important, and using patients’ past experiences and preferences is more important in smoking cessation than in many other areas of medical practice.
One issue to highlight is the fact that the comparator treatment in the trial by Aubin et al1 was the nicotine patch. The patch is the most widely used NRT preparation in the UK13 and elsewhere because it is easy to use. However, in clinical practice it is common for patients to swap forms of NRT. Overall, there is insufficient evidence to suggest that one form is more effective than another,9 but the clinical impression is that allowing someone to renew their confidence by swapping from a choice they see as failing them to another can help.
An important finding from the trial by Aubin et al1 is that the urge to smoke and the occurrence of withdrawal symptoms were reduced with varenicline compared with NRT. Consensus guidelines on measuring and analysing urges and withdrawal suggest doing so primarily in abstinent participants only,14 a practice not adopted by Aubin and colleagues. Data show that urges and withdrawal increase before relapse,15 so any arm that has a higher rate of relapse might have a higher level of urges and withdrawal and therefore these data might only show that resuming smoking was more likely on NRT. Although urges and withdrawal symptoms are associated with relapse, they are probably the cause of most relapses but it is difficult to demonstrate this conclusively. However, Stapleton analysed withdrawal only in largely abstinent smokers and found a similar pattern of results.4 It is likely, then, that the difference in withdrawal is real and that this is an important reason why varenicline is more effective.
It is relevant to ask why a partial nicotinic agonist such as varenicline can achieve greater suppression of nicotine withdrawal symptoms than a full agonist such as NRT. This is probably because the amount of nicotine absorbed from NRT is considerably less than from cigarettes for most smokers,16 and hence they suffer needless withdrawal. There is strong clinical evidence that 4 mg gum is more effective than 2 mg gum in more dependent smokers (85% increase in efficacy), and that adding two or more forms of NRT—most sensibly the patch plus a more rapid acting product (gum, nasal spray, inhalator or lozenge/microtab)—is more effective, increasing the chance of success by 35%.9 The study by Stapleton et al showed that combination NRT was more effective than single form NRT with no evidence that varenicline was more effective than combination NRT.4 However, the study could not exclude a worthwhile difference in efficacy and the point estimate of effect favoured varenicline. The key message is that combining forms of NRT should be the norm in specialist smoking cessation clinics and considered for smokers who may have relapsed on NRT but are keen to use it again in other contexts.
Aubin and colleagues also report that reinforcement from cigarettes smoked while on medication was lower on varenicline than on NRT. However, these data are difficult to interpret. First, most people who sustained prolonged abstinence probably did not smoke at all after quit day if this population is typical. Nothing predicts relapse as strongly as early lapses.17 Thus, this analysis will be dominated by measurements taken at week 1, prior to quitting, when participants were using varenicline but participants randomised to NRT were on no medication. It is unclear whether reduced reinforcement from cigarettes smoked before quitting partly explains the efficacy of varenicline, but positive hedonic responses to cigarettes smoked after quitting are associated with a slight increase in relapse.18 These data suggest that treatment before quitting might be important. In fact, clinical data show that treatment of patients with a nicotine patch for 2 weeks before quitting is associated with a statistically significant increase in abstinence of 79% over post-cessation use only.9 The effectiveness of pre-cessation use is best explained by reductions in reward from smoking weakening conditioned responses, but mediation analysis of the effect of pre-cessation NRT or varenicline on cigarette reward would be needed to confirm this. Whatever the mechanism, pre-cessation use of nicotine patch should be considered in some smokers, although further large trials are probably necessary before pre-cessation use of NRT becomes the norm.
The results of the study by Aubin et al are to be welcomed because varenicline provides another option for smokers wanting to quit. We can now provide clearer statements to our patients that varenicline is the best single pharmacotherapy to assist an attempt to quit. Although varenicline is more effective than the current standard patch, there are good reasons to assume that NRT taken before quit day and in combination might be of similar efficacy. NRT may become even more effective if newer, higher dose, rapid release formulations become available. In my view, however, in varenicline we are close to an optimal treatment for nicotine dependence, but stopping smoking involves more than treating nicotine dependence. Most people who successfully reach the end of treatment will relapse to smoking. While taking longer courses of varenicline may prevent some relapse, lifetime varenicline is unlikely to be a major solution. Instead, we need to develop cognitive-behavioural interventions together with judicious use of various medications. Until then, varenicline is a welcome treatment option that is likely to prove popular with clinicians and patients.
Funding: PA is funded by the National Institute for Health Research (NIHR) but the views expressed are not necessarily those of the NIHR or the NHS.
Competing interests: PA has done consultancy work for the pharmaceutical and biotechnology industries that led to payments to him and his institution. This includes work for companies providing smoking cessation medication including McNeil, who manufacture NRT, and Pfizer, the makers of varenicline.