The role of flavors in attractiveness of electronic cigarettes

Electronic cigarettes (e-cigarettes) are devices that vaporize a liquid (e-liquid), resulting in an aerosol (vapor) that is inhaled by the product’s user. E-liquids consist of a base of propylene glycol and vegetable glycerin, and mostly contain the addictive substance nicotine. E-liquids are available in many different flavors, such as pineapple, chocolate, muffin, mojito, tobacco, and hazelnut. E-cigarette use (i.e., vaping) is less harmful than smoking combustible tobacco, and therefore an attractive alternative for people who aim to quit tobacco smoking. However, e-cigarette emissions contain toxic and addictive compounds. The use of e-cigarettes by people who do not smoke therefore increases their health risks. Although most adult e-cigarette users in the Netherlands are concurrent or former smokers, concerns are raised that e-cigarette use also becomes increasingly popular among adolescents.Of all product characteristics, flavor is most important in the attractiveness of e-cigarettes for both smokers and non-smokers. Flavors in e-cigarettes are currently not regulated on European level. Whereas most e-cigarette users prefer e-liquids with a fruit, sweet, or traditional tobacco or menthol flavor, flavor preferences may differ between user groups. This may offer opportunities for regulation: if e-cigarette flavors could be identified that are attractive to smokers but not to youth and non-smokers, regulators could decide to allow only these e-cigarette flavors. This way, they may be able to facilitate smoking cessation, while preventing the use of e-cigarettes and associated health risks among young people and non-smokers. Therefore, the aim of this thesis was to investigate the role of flavors in attractiveness of e-cigarettes with respect to different user groups.The first sub aim of this thesis was to structure the large amount of e-liquid flavors available. In Chapter 2, we identified and summarized e-liquid flavor names and categories mentioned in literature. We developed a flavor wheel with 16 main categories for consistent categorization of e-liquids based on their marketed flavor descriptions. In order to take regulatory decisions on e-liquid flavors, it is important to gain insight into the flavors that are available. Therefore, in Chapter 3, we created an overview of the flavors of e-liquids marketed in the Netherlands in 2017 by classifying these e-liquids, based on information declared by manufacturers, into the categories of our flavor wheel. We found that the Dutch e-liquid market comprised nearly 20 000 e-liquids in 245 different, mostly sweet, flavors.The second sub aim of this thesis was to identify the most prevalent flavoring ingredients (flavorings) in e-liquids. In Chapter 4, we presented an overview of the flavorings that were most frequently added to e-liquids, in general and per flavor category, using information declared by manufacturers in 2017. We found that manufacturers most often add vanillin (sweet, vanilla-like flavor), ethyl maltol (sweet, fruity-caramellic flavor) and ethyl butyrate (ethereal, fruity flavor) to their e-liquids, and we identified 29 flavorings that were specific to a respective flavor category. Based on the similarities and differences in e-liquid flavoring compositions between flavor categories, we could predict e-liquids’ flavor categories with 70% accuracy using a machine learning algorithm. As data from manufacturers are not always complete and correct, we identified e-liquid flavorings in Chapter 5, using chemical analysis of 320 e-liquids classified in various flavor categories. The flavorings detected most often were vanillin, ethyl butyrate, and cis-3-hexenol (fresh, grassy flavor). In addition, we found that flavoring compositions of e-liquids were similar within fresh/sweet, warm/sweet, fresh/cooling, and non-sweet flavor categories, respectively.The third sub aim of this thesis was to determine which flavors (potential) users of e-cigarettes prefer and like the most. In Chapter 6, we conducted survey research and found that, in the Netherlands, smokers were mostly interested in e-cigarettes with a tobacco or menthol/mint flavor, whereas people who had never smoked nor used e-cigarettes were mostly interested in sweet and menthol/mint flavors. Sweet and fruit flavors were also reported to be regularly used by people who use e-cigarettes (and do not smoke). As survey research is based on participants’ memory and mental representation of how they perceive a particular flavor, this is a more indirect approach to investigating flavor preferences compared to sensory research, during which participants can actually taste or smell the product. As it is unethical to expose non-smokers and adolescents to e-cigarette emissions, research in these groups requires an alternative approach. In Chapter 7, we aimed to determine whether smelling could be an alternative to vaping in the hedonic assessment of e-cigarette flavors. We found a strong correlation between smelling and vaping for the liking of e-liquid flavors, that did not differ between smokers and non-smokers. Finally, in Chapter 8, we investigated sensory liking of various tobacco and non-tobacco e-liquid flavors in adolescent non-smokers, young adult non-smokers, and adult smokers. We found that both sweet and menthol/mint e-liquid flavors were liked similarly across all groups, and that these flavors were clearly liked more than tobacco flavors.In general, these findings consistently show a clear preference for sweet and menthol/mint e-liquid flavors among all groups of (potential) e-cigarette users. These flavors contribute to liking and appeal of e-cigarettes directly by, respectively, enhancing sweetness and producing cooling sensations, and indirectly by masking the bitter taste and harsh/irritating sensation from nicotine. Our findings suggest that banning all e-cigarette flavors except tobacco will reduce attractiveness of e-cigarettes for all (potential) user groups, including adolescent and young adult non-smokers as well as adult smokers. Such a ban, as recently announced in the Netherlands, could be based on (1) the flavor as mentioned on the product label, (2) the flavor as perceived by users of the product, and/or (3) the flavoring ingredients in e-liquids that result in the perceived flavor. This could be enforced by analyzing product information declared by manufacturers in combination with chemical-analytical and/or sensory data.