E-cigarettes are some young people's training wheels for smoking. Here's the evidence

I started on mother’s milk, soon graduated to cow’s milk, went onto fizzy drinks, took a liking to beer, moved with gusto onto wine, and in my 30s settled for single malt. So, did my mother’s breast milk lead me to Scotch whisky? Was it a “gateway” to all the devil’s best tunes later in life?

The “gateway hypothesis” in some areas has often taken a deserved shellacking for its imprecision. It is an argument most often wheeled out over illicit drugs: a toke on a joint will be soon followed by injecting heroin.

Today the gateway hypothesis is being given a huge workout over concerns e-cigarettes might lead to kids smoking.

A series of nine recent studies has shown young people who try e-cigarettes are more likely to go on to smoke cigarettes by the next time they are interviewed. These studies all considered young people who had not smoked a cigarette, and then compared smoking between those who did, and did not use e-cigarettes at the start.

E-cigarette advocates have often dismissed these studies by saying all they show is that “children who are going to smoke in the future, will smoke in the future” and “kids who try, stuff, try stuff”.

Here they are alluding to the very real issue that some children are likely to have a constellation of existing vulnerability factors that make them more likely to use e-cigarettes and smoke.

Researchers try to tease this out by using measures of this vulnerability and then seeing whether there’s still an association between using e-cigarettes and later smoking even after these factors have been taken into account.

Predictors of smoking (not including e-cigarette use) include general risk-taking, impulsiveness, low self-esteem, parental smoking, and affiliation with other risk-taking peers.

What do these studies show?

For the first time, researchers have combined the results of these nine studies, involving 17,389 people aged 14-30, and analysed the results (in a type of analysis known as a meta-analysis).

They found e-cigarette users were nearly five times more likely to smoke than those who had not used e-cigarettes. But this was reduced to a three-fold increased risk after adjusting for demographic, psychosocial and behavioural risk factors that predict cigarette smoking.

So, even after taking into account the very factors e-cigarette advocates argue would muddy the evidence, there was a three-fold increased risk of young e-cigarette users going on to smoke.

Yet, when faced with the argument that e-cigarettes may act as a catalyst to subsequent smoking, e-cigarette advocates disagree.

Typical responses contrast with what we know about the relationship between early influences on later behaviours, across a vast range of health and social problems.

For example, we search for factors that might lead to some falling in with terrorists. We try to understand the antecedents of obesity.

We know low price, tobacco advertising, parental smoking, smoking by best friends, attractive packaging, and weak, non-explicit health warnings and awareness campaigns are all associated with higher rates of smoking in some groups. This understanding has informed policy and practice and we now have a blueprint for comprehensive, effective efforts that have combined to drive teenage smoking down to its lowest ever levels.

But when research suggests using e-cigarettes might condition some children into thinking, “I wonder what the ‘real thing’ [ie smoking] is like?”, that idea is completely out-of-bounds, apparently.

When UK researcher Peter Hajek, a long-time advocate of e-cigarettes, commented on the results of this meta-analyis, he said:

People who drink white wine are more likely to also try red wine than teetotallers, but common sense would not suggest that this means that removing the white will reduce the drinking of the red.

But this is a poor analogy, as one of the authors of the JAMA meta-analysis suggests:

Young people report that there is a lot of pressure among e-cigarette only users to smoke a “real” cigarette. It may be somewhat analogous to the fact that teens who use flavoured alcohol are often pressured socially to step up their game to harder forms of alcohol.

E-cigarettes, with their many teen-friendly flavours, their far less harsh “throat grab”, the ease with which they can be used inconspicuously (little smell, easily hidden), and their “almost totally safe” pitch have considerable appeal to young people compared with smoking.

The ‘vaping rising, smoking falling’ argument

Another argument that advocates of e-cigarettes use to argue against e-cigarettes acting as “gateway” products is that in young people, vaping is rising and smoking is falling. If they acted as “gateways” to smoking, they argue, surely we’d see smoking rates in young people rising.

But this does not follow at all.

There are multiple reasons for both the rise and the fall in smoking. These include price changes, the denormalisation and growing social unacceptability of smoking, anti-smoking campaigns, the impact of advertising bans (you’d have to be over 24 in Australia to have ever seen a local tobacco advertisement after it finished here in 1993).

You can still have fewer young people smoking – teenage smoking is down to its lowest ever levels in Australia – but still be having a “gateway” effect.

If the impact of all factors driving smoking down in young people was greater than the impact of any “gateway” effect to smoking, the figures would still show fewer young people were smoking, which is what’s happening: there’s a net fall.

If the fall in young people smoking is so great, it could mask the any rise in smoking caused by any e-cigarette gateway effects, which could still be substantial.

For this reason, the types of studies pooled in the JAMA meta-analysis are critical in understanding whether e-cigarettes are an important catalyst for young people to smoke.