Comprehensive Overview of Common E-Liquid Ingredients and How They Can Be Used to Predict an E-Liquid’S Flavor Category
Total Page：16
File Type：pdf, Size：1020Kb
Original research Tob Control: first published as 10.1136/tobaccocontrol-2019-055447 on 10 February 2020. Downloaded from Comprehensive overview of common e- liquid ingredients and how they can be used to predict an e- liquid’s flavour category Erna J Z Krüsemann ,1,2 Anne Havermans ,1 Jeroen L A Pennings,1 Kees de Graaf,2 Sanne Boesveldt,2 Reinskje Talhout1 10–14 ► Additional material is ABSTRact cigarette smoking. In line with this, the use and published online only. To view Objectives Flavours increase e- cigarette attractiveness marketing of e- liquid flavours that are appealing to please visit the journal online (http:// dx. doi. org/ 10. 1136/ and use and thereby exposure to potentially toxic smokers may contribute to public health benefits. tobaccocontrol- 2019- 055447). ingredients. An overview of e- liquid ingredients is needed However, flavours may also stimulate vaping among to select target ingredients for chemical analytical and non- users, in particular young people.15–17 This is 1 Centre for Health toxicological research and for regulatory approaches concerning, as e- cigarettes are not safe10 18 19 That Protection, Rijksinstituut voor aimed at reducing e- cigarette attractiveness. Using is, chemicals in e- cigarette emissions (eg, tobacco- Volksgezondheid en Milieu, Bilthoven, The Netherlands information from e- cigarette manufacturers, we aim to specific nitrosamines, metals, aldehydes and other 2Division of Human Nutrition identify the flavouring ingredients most frequently added flavourings) can be toxic and thus harmful to and Health, Wageningen to e- liquids on the Dutch market. Additionally, we used consumers’ health.20–22 In addition, e- cigarettes University, Wageningen, The flavouring compositions to automatically classify e-liquids may facilitate smoking initiation among never- Netherlands into flavour categories, thereby generating an overview smokers.23 As a consequence, e- liquid flavours are that can facilitate market surveillance. Correspondence to considered an important target in tobacco control Methods We used a dataset containing 16 839 Erna J Z Krüsemann, Centre for in order to decrease e-cigarette attractiveness and Bibliotheek. Protected by copyright. Health Protection, Rijksinstituut e- liquids that were manually classified into 16 flavour use and thereby exposure to potentially toxic voor Volksgezondheid en categories in our previous study. For the overall set and emissions. Milieu, Bilthoven 3721 MA, each flavour category, we identified flavourings present Netherlands; One way of regulating e-liquid flavours could be erna. krusemann@ rivm. nl in more than 10% of the products and their median restriction of flavour categories that are particularly quantities. Next, quantitative and qualitative ingredient appealing to non- users or youth. For example, the EJZK and AH contributed information was used to predict e- liquid flavour US Food and Drug Administration (FDA) recently equally. categories using a random forest algorithm. announced that e-cigarettes with a flavour other Results We identified 219 unique ingredients that were Received 10 October 2019 than tobacco will be removed from the market Revised 10 January 2020 added to more than 100 e- liquids, of which 213 were until they are evaluated under the FDA’s new Accepted 19 January 2020 flavourings. The mean number of flavourings per e- liquid product review authority.24 Such rules often rely on Published Online First was 10±15. The most frequently used flavourings were flavour-related marketing descriptions, which do 10 February 2020 vanillin (present in 35% of all liquids), ethyl maltol (32%) not necessarily reflect human perception and may and ethyl butyrate (28%). In addition, we identified 29 differ between e-liquid providers and therefore can category- specific flavourings. Moreover, e- liquids’ flavour be challenging to enforce. In addition, allowing or http://tobaccocontrol.bmj.com/ categories were predicted with an overall accuracy of banning a specific e- liquid flavour may be compli- 70%. cated as e- liquids regularly have multiple flavour Conclusions Information from manufacturers can be descriptors (eg, tobacco with caramel and vanilla).7 used to identify frequently used and category- specific Similarly, using sensory analysis to enforce e- liquid flavourings. Qualitative and quantitative ingredient flavour regulations may be challenging. Although information can be used to successfully predict an based on human perception, analysing sensory e- liquid’s flavour category, serving as an example for properties of all e- liquids in a particular country regulators that have similar datasets available. to determine whether they comply with current regulations is extremely time-consuming. However, sensory analysis could be used to determine attrac- on March 26, 2021 at Wageningen UR INTRODUCTION tiveness of particular e- liquid flavours or flavour Electronic cigarette (e-cigarette) use among various categories among vulnerable user groups (eg, user groups has increased considerably over the non- smoking adolescents) and in that way inform past years.1 2 One of the most important reasons policy makers on how to reduce e-liquid appeal. for e- cigarettes’ great popularity is the assortment Another option would be to decrease e- liquid © Author(s) (or their of available e-liquid flavours3–6; for example, no attractiveness or toxicity by restricting the addition employer(s)) 2021. Re- use less than 245 unique flavours were available in the of particular flavouring ingredients. For example, permitted under CC BY- NC. No 7 commercial re- use. See rights Netherlands in 2017. Flavours increase product particular flavourings could be banned or their and permissions. Published attractiveness among all types of (potential) users, maximum concentration could be restricted. This by BMJ. that is, among youth and adults8 and among may diminish e- cigarette use and thereby exposure To cite: Krüsemann EJZ, current smokers, dual users, exclusive vapers, as to potentially toxic e- liquid constituents and thus 9 Havermans A, Pennings JLA, well as never- users. For smokers, switching to increase health benefits for non-users and youth. et al. Tob Control e- cigarettes may be beneficial, as e- cigarette use Research on e- liquid flavours and flavouring 2021;30:185–191. (ie, vaping) is considered less harmful than regular compositions can support regulators in developing Krüsemann EJZ, et al. Tob Control 2021;30:185–191. doi:10.1136/tobaccocontrol-2019-055447 185 Original research Tob Control: first published as 10.1136/tobaccocontrol-2019-055447 on 10 February 2020. Downloaded from policy measures. Accordingly, various chemical analytical studies Data analyses evaluated e-liquid flavouring ingredients and emissions.21 25–29 As most e- liquids were reported as being marketed in a package These studies typically focus on a list of a priori selected target unit of ‘1’ and containing 10 mL of fluid, deviating submissions flavourings for their analyses.21 25–29 These target lists are usually (more than one e- liquid per package or e- liquids with a volume selected based on previous studies, which creates a risk of selec- other than 10 mL) were excluded. This resulted in a final dataset tion bias by overlooking other or new e-liquid ingredients that of 16 839 products. For these products, ingredient- related infor- may have toxic properties. Therefore, there is a clear need for mation was extracted from the EU- CEG system and analysed a published overview of common flavourings and other ingredi- using R statistical software (version 3.5.1). Ingredients reported ents in e- liquids. This information may be used as a starting point by manufacturers as having the function ‘Flavour and/or Taste for future chemical analytical researchers in developing their Enhancer’ will be referred to as flavourings. Negative values for target lists, may provide targets for future toxicological studies ingredient amounts (resulting from EU- CEG artefacts) were set and may provide foci for regulation of e- liquid flavourings. at zero. For the overall dataset, as well as for individual flavour We previously generated an overview of all e- liquid flavour categories and the unclassifiable subset separately, the following descriptions reported to be marketed in the Netherlands in values were determined: the number of products, the mean 7 2017, by manually classifying almost 20 000 e- liquids into 16 number of total ingredients, the mean number of flavourings per 30 main flavour categories. Classification was based on e-liquids’ product, the mean total number of ingredients per product and flavour-related information reported by manufacturers. In the the mean quantities of all flavourings per product. European Union, manufacturers are also required to provide Ingredients present under multiple names in the EU- CEG information about their e-liquids’ chemical composition. Using system (eg, ethanol, etanolo, etanol, ethyl alcohol, ethyl alkohol, this information, the current study first aims to identify the most ethyl- alcohol, alcool ethylique and EtOH) were merged into one commonly used e-liquid flavourings in general and to determine ingredient name. First, unique ingredient names were identified potential flavourings that are specific to a single flavour category. (n=8352), including the number of products for which they were In our previous study, we used information from manufac- reported in the EU- CEG system. Next, starting with the most turers such as brand names to manually