Monday, May 16, 2005

Anti-Smoking Vaccine Hype Vastly Overblown

A vaccine against nicotine is being touted as a breakthrough in some enthusiastic media coverage, but a more cautious analysis reveals the new treatment falls far short of the apparent success.

The Phase II trial randomly assigned 341 moderate-to-heavy smokers either the vaccine or a placebo. Participants got five injections over five months and were encouraged to stop smoking.

Here's how the Times of London, broke the news :
"A vaccine against nicotine has produced impressive results in a large-scale human trial. The success of the vaccine was predicted in The Times. Almost 60 per cent of those who responded best succeeded in giving up smoking for at least 16 weeks. Those who did not succeed in giving up smoked significantly less. Among those who responded less strongly, a third managed to quit, an encouraging result but not significantly better than in those given a placebo — just under a third of whom also managed to give up." [source]
The results for Cytos002-NicQb vaccine sound promising. But there are lies, damm lies and pop-medicine statistics. Take a close look at this Times graphic:

Notice that no combination of the percentages adds up to 100%.

A casual glance implies that 57% of those vaaccinated gave up smoking. But actually this represents the 57% of high antobody responders who ceased. Read that Times intro again: "Almost 60 per cent of those who responded best succeeded in giving up."

That's a sentence it pays to read slowly.

In fact, there was no statistically significant difference in cessation rate between the vaccine group and those taking a placebo. Cessation was 57% among the already less than half of smokers who had developed a high level of antibodies. For the rest, the five month course of vaccinations performed as poorly as placebo. [data]

Here's the statistical trickery: Lets say 51% of women who took a new contraceptive pill had no pregnancies, but the other 49% did get pregnant. Not very impressive. But, by a similar logic I could now claim that 98% of those who responded well did not get pregnant. But you might be a little upset if you found you had been conned by my statistic.

Another issue is that the medical setting and circumstance of the vaccinations would have ensured motivation was high for participants getting both real and fake vaccine. All participants received smoking cessation counselling by the study collaborators at each visit during three months. Even the Times admits the motivation issue skewed the results:
"Surprisingly, 31 per cent of the placebo group were also able to stop smoking, a much higher figure than would be expected normally. This shows the value of counselling combined with the belief that they were being given an effective treatment."
This also shows that suggesting to people they will loose their addiction is a powerful therapy in itself -as the limited success of many quit-smoking therapies attests. So this effect should really be subtracted from the claimed beneficial results.

And all benefits -real or imagined- should be offset against the medical risks of such therapy. The vaccine is based on an altered protein from a type of virus that attacks bacteria. Patients develop antibodies against nicotine. Side-effects included flu-like symptoms on the day of injection, redness and swelling at the injection site.

The new smoking vaccine is the start of a trend. Dr. Wolfgang Renner, Chief Executive Officer of Cytos Biotechnology, says:
"This finding is extremely important with respect to our pipeline of 27 vaccine candidates in other major disease areas like high blood pressure, obesity or Alzheimer's disease where we use the same basic... principle."
Let's hope the reporting is more skeptical as this technology moves on to new applications.

Now that I'm finished writing this, maybe I should have a cigarette. Just for the health benefits, you see. Oh, hadn't you heard? It seems that 100% of non-smokers eventually die. Chilling statistic eh?

Don't become a number. Start smoking today.


Post a Comment

<< Home