The "Gateway" hypothesis: evaluation of evidence and alternative explanations.

Background: Electronic nicotine delivery systems (ENDS) offer a substantial harm reduction opportunity for adults who smoke and are unlikely to quit. However, a major concern about ENDS is their use by non-smoking youth, and particularly whether ENDS are acting as a "gateway" that leads youth to later start smoking cigarettes. However, evidence for the gateway hypothesis can be interpreted in alternative ways, e.g. that youth who have certain characteristics were already predisposed to use both ENDS and cigarettes ("common liability" explanation).

Aims: This commentary provides an evaluation of the gateway hypothesis that is accessible by a lay audience. This paper first reviews and evaluates the evidence interpreted as supporting the gateway hypothesis. Important alternative explanations (i.e., common liability) are discussed, as are different types of evidence (i.e., population-level trends) that can help differentiate between these competing explanations.

Overview: Evidence for the gateway hypothesis is based on the finding that youth who use ENDS are more likely to also smoke cigarettes. However, this evidence suffers from an important flaw: these studies fail to fully account for some youths' pre-existing tendency to use products containing nicotine, and inappropriately interpret the results as ENDS use causing some youth to smoke. Common liability studies suggest that ENDS use does not, in and of itself, directly cause youth to later smoke cigarettes, beyond their pre-existing tendency to use products containing nicotine. Population-level trends show that youth cigarette smoking declined faster after ENDS use became common, which contradicts the central prediction of the gateway hypothesis (i.e. that youth smoking would be more common following ENDS uptake, than otherwise be expected).

Conclusion: Evidence offered in support of the gateway hypothesis does not establish that ENDS use causes youth to also smoke cigarettes. Instead, this evidence is better interpreted as resulting from a common liability to use both ENDS and cigarettes. Population-level trends are inconsistent with the gateway hypothesis, and instead are consistent with (but do not prove) ENDS displacing cigarettes. Policies based on misinterpreting a causal gateway effect may be ineffective at best, and risk the negative unintended consequence of increased cigarette smoking.