Observations Yankee Doodling

Does nicotine replacement really help smokers quit?

BMJ 2012; 344 doi: http://dx.doi.org/10.1136/bmj.e450 (Published 18 January 2012) Cite this as: BMJ 2012;344:e450
  1. Douglas Kamerow, chief scientist, RTI International, and associate editor, BMJ
  1. dkamerow{at}rti.org

New study throws doubt on a prevention mainstay

As I write this, a front page story in the New York Times1 is reporting a population based study that followed smokers who had recently quit.2 The takeaway message is that “the nicotine gum and patches that millions of smokers use to help kick their habit have no lasting benefit and may backfire in some cases.”1

Wow, really? I’m all for evidence based medicine and evidence based public health (and evidence based everything, I guess), but is this conclusion really justified?

Before we take a look at the study, a few background notes. Tobacco use remains the leading cause of death in the United States, responsible for one in five deaths, around 440 000 a year.3 Smoking is addictive; most smokers are dependent on nicotine.4 Nicotine replacement therapy, introduced in 1984, is now available in multiple forms: gum, lozenge, patch, inhaler, and nasal spray, the first three of which are sold in the US without a doctor’s prescription. A huge number of well done randomised controlled trials and meta-analyses in the clinical setting have shown that nicotine replacement helps smokers quit in the short term and to stay off tobacco for up to 12 months.5 About a third of US smokers who try to quit use over the counter products.6

The study that has grabbed the headlines is a prospective cohort study done in Massachusetts.2 A total of 787 smokers who had recently quit were interviewed in three waves over six years. They were asked about quitting, relapsing, and their use of nicotine replacement and professional counselling to help them quit. Slightly more than 20% reported using nicotine replacement to help them quit. About 30% of patients in each wave had relapsed and were smoking again. The big finding of the study was that those who used nicotine replacement with or without counselling relapsed at roughly the same rate as those who did not.

The authors concluded that their findings “raise serious questions regarding the population level effectiveness” of nicotine replacement and counselling. Further, they are concerned that “funding ineffective services that aim to change individual behaviour may be resulting in the loss of scarce resources from public health programmes” such as mass media campaigns, tobacco tax rises, and antismoking regulation.

Are these conclusions justified? Can a single cohort study nullify the results of a hundred randomised controlled trials? The drawbacks of randomised controlled trials are well known. Although they excel at excluding most sources of bias, they are almost always performed in research settings among homogeneous populations and exclude (by design) many population groups. Studies of nicotine replacement are usually done on heavy smokers, for instance. Population studies such as this one, performed in the real world, presumably include a much more diverse group. This should be especially useful when examining an over the counter product that is often used without a doctor’s supervision. Thus, we should welcome well done population based studies of these issues.

But was this study well done? There is plenty to quibble about. Firstly, from the paper it is hard to tell exactly what the actual response rate was. The authors state it is 46%, but they started with a probability sample of 6739 adults and ended up with a study sample of 787 who reported recently quitting at wave 1 or 2 and were then interviewed in the next wave. There is no way to know whether the ultimate participants were similar to the non-participants. Of the 787, only one in five actually used nicotine replacement to help them quit, so now we’re down to about 165 or so. Only 30% relapsed, so the whole analysis revolves around about 50 relapsers who used nicotine replacement. Notably, just a third of them said they used the treatment for the recommended eight weeks, leaving about 16. So, not much to hang grand conclusions on.

Secondly, there is recall bias. People who seek treatment, whether this is counselling or drugs, are likely to remember it more than if they make a less committed, more informal, attempt to quit without these treatments.6 This distorts results of surveys that are based on recall.

Finally, and most importantly, we have known for a long time that people who quit on their own (“cold turkey”) are different from those who need and seek help through counselling or drugs. Most smokers who quit do so without any assistance.7 They may be less addicted, or have more willpower, or something else, but it is certainly no surprise to learn that an unselected group of relapsing smokers will have no greater proportion quitting unassisted than by using nicotine replacement therapy.

For all these reasons, it is disappointing to see the brouhaha that this study has caused. I don’t think there is any doubt that nicotine replacement helps smokers quit and stay off tobacco products for at least a year when done in the clinical setting. There is some evidence in use of over the counter products as well, but we clearly could use more good studies in the community to guide recommendations and more long term follow-up studies as well.

In addition, it really makes no sense to imply that money is being diverted from public health programmes and wasted on nicotine replacement. To effectively fight tobacco use we need a combination of strategies: clinical, public health, regulatory, and legislative. It is not a zero sum game.

Notes

Cite this as: BMJ 2012;344:e450

Footnotes

References