Synthetic cooling agents in tobacco and other consumer products: history and chemistry

Menthol, mint oils, menthone and associated chemicals are prominent flavour compounds added to body care, food and pharmaceutical products. Though widely used for its cooling properties, menthol, at higher concentrations, produces intense irritation to mucous membranes.33 34 Menthol’s high volatility may also cause deterioration of the freshness of mentholated tobacco products during storage.35 36 To circumvent these drawbacks, industry pursued research and development of chemicals that possess menthol-like cooling activity but with low volatility and without any minty aroma. Using structure-activity relationship approaches, WS in the 1970s generated a library of 1200 synthesised cooling agents.11 37 The agents were labelled using the moniker WS, the manufacturing company’s initials and an identifying number. Some of these coolants were based on the p-menthane structure, the basic monoterpene structure of menthol, with derivatives modified with polar groups directly attached to p-menthane via H-bonding (eg, WS-3, WS-5). Several acyclic compounds that possessed p-menthane-like structural elements were found to retain cooling activity (eg, WS-23).11 Similar to menthol isomers, different synthetic cooling agent stereoisomers or racemic mixtures varied in their cooling activity, odour and taste.11 12 37

Our search of the Truth Tobacco Industry Documents archive found evidence that WS collaborated with RJ Reynolds in 1974–1975 to develop a new tobacco product class, Cool-Without-Menthol, by incorporating non-menthol cooling substances into tobacco products.38 Industry documents indicate that most of the synthetic coolants in consumer market testing did not increase user appeal, except for tobacco products with WS-14, which never were marketed publicly.38 In 1981, Philip Morris distributed cigarettes containing WS-14 as ‘Northwind Chill Blend’ in US test markets, with advertising themes such as ‘NORTHWIND…cools like no menthol can’.39 40 Industry documents indicated that insufficient knowledge regarding WS-14’s health effects in combusted tobacco smoke prompted Philip Morris’ discontinuation of Northwind cigarettes.

The Flavour and Extracts Manufacturers Association (FEMA), a flavour industry organisation, eventually granted Generally Recognized as Safe (GRAS) status to several synthetic coolants (eg, WS-3, WS-5, WS-12, WS-23, menthyl lactate, Frescolat-ML).11 FEMA GRAS status is limited to the intended use of a given chemical in food flavourings. The FEMA GRAS programme has not always provided safety assessment data with the Food and Drug Administration (FDA) or expedited public release of data. Consequently, it is difficult to independently verify manufacturer safety claims and identify every compound that does not receive marketing authorisation or GRAS status. Currently marketed products containing synthetic cooling agents include chewing gum, confectionary products, shaving cream, beverages, and body and dental care products. FEMA GRAS status does not extend to inhalational safety. Nevertheless, an analysis detected small but significant amounts of WS-23 in combustible cigarettes sold in Germany.10 41

In summary, the tobacco industry has conducted research on and market-tested synthetic cooling agents for 40 years. Additionally, the industry has marketed some combustible tobacco products that contain synthetic coolants.

Non-menthol cooling agents in e-cigarettes

One study measured a significant concentration of WS-3 (~0.3%)8 10 in mint-flavoured European JUUL pods. Another study detected WS-23 in US-marketed JUUL pods marketed as menthol-flavoured at an average concentration of 0.1 mg/mL (~0.01%).42 This study also investigated US-marketed PuffBar disposable e-cigarettes, detecting high WS-23 concentrations in PuffBars marketed as mint-flavoured (mean: 36 mg/mL,~3.6%).

Jabba et al analysed e-liquid flavour constituents in 14 different PuffBar brand disposable e-cigarette devices (6 marketed as fruit-ice hybrid flavours (e.g. ‘blueberry ice’), 1 cool mint, 1 menthol, 5 fruit-only (e.g., ‘strawberry-banana’) and 2 other (eg, ‘café latte’)) purchased in 2020.9 Only one PuffBar product (ie, marketed as ‘pomegranate’ flavoured) did not contain WS-3 or WS-23. Almost all flavours contained WS-23 and several flavours contained a combination of two or three cooling agents (WS-3, WS-23 and menthol). PuffBar flavours marketed as either ‘Ice’, ‘cool mint’, ‘menthol’ or fruit +ice flavour combinations had high concentrations of WS-23. Even some PuffBar flavours marketed as flavours that did not connote cooling (eg, ‘Lychee’ flavour) had high WS-23 concentrations. This same study tested e-cigarette refill solutions marketed with the terms ‘koolada’, ‘kool/cool’, ‘ice’ or containing WS-3/WS-23 (eg, ‘Popsicle USA’, ‘frozen berries’, ‘ice dragon’). WS-3 was detected in nearly all of the 25 refill solutions and half contained WS-3 and menthol in combination.9

These studies provide converging evidence for a diversity of non-menthol cooling agents in various e-cigarette products. They also demonstrate that the characterising flavours on product labels often do not include monikers referring to cooling effects or constituents, as in the example of the abovementioned fruit or berry flavours.

Pharmacology of synthetic cooling agents

Similar to menthol, WS-3, WS-23 and Frescolat-ML have pharmacological activity at TRPM8 (transient receptor potential cation channel subfamily M member 8), which is the cold/menthol receptor expressed in sensory nerve endings innervating the oral cavity and upper airways.25 43–45 TRPM8 mediates menthol’s cooling, counterirritant, analgesic, antitussive, anaesthetic and anti-inflammatory effects that suppress irritation from nicotine and other irritants in tobacco product aerosol.25 43–45

The half-maximal effective concentrations (EC₅₀) for WS-3, WS-23 and Frescolat-ML to activate TRPM8 are 3.7 µM, 44 µM and 3.3 µM, respectively, and are comparable to l-menthol’s EC₅₀ (4 µM).11 25 43 46 Compared with l-menthol, WS-3 has higher efficacy (maximum-response) at TRPM8, producing stronger perceived cooling sensations (150%).11 47 Some synthetic coolants also activate the sensory irritant receptor, TRPA1 (transient receptor potential cation channel subfamily A member 1; acrolein or wasabi receptor), although with lower potency or efficacy than menthol.46 48 49 (Sven Eric Jordt and Sairam V Jabba, unpublished results) Synthetic coolants in e-cigarette liquids have high carryover rates into aerosol (91-≥100%) similar to menthol, indicating that coolants reach users’ airways.9 50

In sum, synthetic coolants as compared with menthol have comparable or stronger pharmacological activity at TRPM8, cause less irritation, and might elicit stronger cooling and counterirritant effects. We speculate that these properties may lead to diminished sensory irritation and nicotine aversion, which could potentiate appeal of the user experience, promote deeper aerosol inhalation, and, in turn, encourage more frequent e-cigarette use. More intensive puffing patterns encouraged by the anti-irritant effects of synthetic coolant could presumably lead to higher temperatures produced by heating coils, which could result in a greater quantity and quality of toxins in aerosol.

Toxicity of synthetic coolants

A review of unpublished documents by the Joint Expert Committee on Food Additives (JECFA) of the Food and Agriculture Organisation of the United Nations, and the WHO indicate a threshold of toxicological concern of 90 µg/person/day (1.5 µg/kg body weight/day) for WS-3 and WS-23 ingestion.51 This threshold value is common to all chemicals included under structural class III to which WS-3 and WS-23 belong.52–56 Thresholds of toxicological concern are based on the premise that a low level of exposure with a negligible risk can be identified for many chemicals within a structural class that share common chemical structures. For each structural class, the threshold of toxicological concern is extrapolated from hundreds of compounds with established toxicity data. From these data, the fifth percentile no-observed-adverse-effect levels in mg/kg body weight per day is converted to intake for a 60 kg person followed by the application of a 100-fold uncertainty factor to calculate the threshold of toxicological concern value.52–56 This approach to identifying thresholds of toxicological concern has been applied by the European Medicines Evaluation Agency, the WHO International Programme on Chemical Safety for the risk assessment of chemicals, and by the EU Scientific Committee on Toxicology, Ecotoxicology, and the Environment.52–56

JECFA documentation on WS-3 and W-23 describes several oral toxicity studies for the compounds,51 53 including a 28-day WS-3 oral toxicity study conducted in rats. That study found mild hepatic and nephrotoxic lesions at >8 mg/kg body weight/day doses.52 53 Rats also exhibited weight reduction and altered haematological and clinical biochemical parameters at higher doses. In a separate 14-week oral toxicity study of WS-23, rats dosed daily demonstrated treatment-related lesions in kidneys (renal tubular nephrosis) and liver (fatty degeneration, vacuolar changes) at doses exceeding 5 mg/kg body weight/day.53 57 Cellular genotoxic assays observed clastogenic effects of WS-23, inducing breaks in chromosomes on metabolic activation.53 A 4-week WS-23 rat inhalation toxicity study conducted by an e-cigarette company (RLX Technology, China) that administered 19 mg/kg body weight/day and 29 mg/kg body weight/day detected exposure-related changes for several lung function parameters.58 Across these studies, no-observed-adverse-effect levels for WS-23 and WS-3 were determined to be 5 mg/kg body weight/day and 8 mg/kg body weight/day, respectively.

Taking into account the concentration of WS-3 and WS-23 detected in various e-cigarette products, one study estimated that vaping 1 mL of e-liquid or one PuffBar per day would exceed the threshold of toxicological concern dose set by JECFA for WS-3 and WS-23 or other structural class III flavour amides (90 µg/person/day) for ingestion.9 That study also estimated that the JECFA flavour amides ingestion threshold would be exceeded by a range of 2-fold to 180-fold for WS-3 and by 26-fold to 598-fold for WS-23.9 51 For these estimations, the expected exposure dose for a user is similar to the amount of synthetic coolant present in the volume of e-liquid that the user vapes as the average carryover efficiency of coolants from e-liquids to aerosol was measured to be ~100%.9 51 In addition, for several of the coolant-containing refill liquids and disposable e-cigarette varieties, predicted margins of exposure (no-observed-adverse-effect-level based risk assessment parameter) were lower than the recommended safety margin of 100 for organ toxicity end points when 1–3 mL of e-liquid (or half a Puff Bar device) is used per day.9 Organotoxic or non-genotoxic chemicals with margins of exposure below the recommended safety margin of 100 raise concern for regulatory agencies to prioritise them for risk-mitigation efforts.

Calculating margins of exposure for inhalation of a compound based on past oral toxicity research requires caution. The systemic risk for synthetic cooling agents may vary across oral and inhalational exposure. It is a common practice among regulatory agencies, such as the European Chemical Agency, to use a factor of 2 or higher for extrapolation of systemic effects from oral-to-inhalation routes.59 This value is based on a default of two to three times higher absorption efficiencies of inhalation versus oral administration that has been observed in a review of results from 246 toxicity studies for 110 different chemicals.60 Hence, a lower chemical exposure via inhalation route may be sufficient to produce similar adverse effects as exposures via the oral route for most compounds.59–61 Studies measuring delivery, retention and pharmacokinetics of various e-cigarette constituents revealed that, on average ~95% of the propylene glycol, vegetable glycerin and nicotine delivered from e-cigarettes are systemically retained by users.62 63

In sum, initial toxicological research suggests that the levels of synthetic cooling agents inhaled from e-cigarettes are sufficiently high to exert both physiological effects and possible toxicological effects. The dose of synthetic coolants that regular users of ice-flavoured products inhale may surpass recommended safety levels set by regulatory agencies for food consumption; however, further researcher is needed to understand and extrapolate toxicity thresholds across inhalation and ingestion of synthetic coolants. Regardless, synthetic coolant containing e-cigarette products merit thorough safety review by regulatory agencies.7 9

Ice-hybrid flavors: potential pharmacological and toxicological implications

Several flavour chemicals identified in fruit and dessert-flavoured e-cigarette products, including benzaldehyde (cherry/berry), vanillin, ethyl vanillin (dessert or sweet), citral, or limonene (citrus) and their flavor-solvent adducts (flavor-aldehyde propylene glycol/vegetable glycerin acetals), activate the irritant receptors TRPA1, TRPV1 and TRPV3.64–67 Addition of selective TRPM8-activating synthetic coolants to these fruit and dessert-flavour constituents might counter the respiratory irritancy of fruity/sweet flavorants as well as provide cooling sensations. The counterirritant effects of synthetic coolants might be similar to the effects gained by the industry shift towards adding organic acids to e-cigarettes to create nicotine salt products, an evolution that reduced the harshness of e-cigarette products with high nicotine concentrations.68

The toxicological implications of inhaling multiple flavour chemicals simultaneously, either in presence or absence of menthol or synthetic coolants, is poorly understood. Research on whether there are additive or synergistic toxicological properties of simultaneous exposure to cooling agents and fruity, minty or other flavouring constituents is needed to inform regulatory agencies of their health risks.

Ice flavours marketed directly to consumers as ‘ready for use’ products

Articles characterising PuffBar brand and other pod-style disposable e-cigarettes illustrate that ice-fruit hybrid flavours were marketed as early as 2019.69 70 Our April 2021 internet search of English-language e-cigarette marketing websites found that ice-flavoured products are marketed in numerous pod-style cartridge, disposable and refillable liquid e-cigarette products (see images of example of ice-flavoured products being marketed online in figures 1 and 2). Ice-flavoured products were identified on both US and UK manufacturer websites in either nicotine salt or freebase formulations at various nicotine concentrations. To examine the availability of ice-flavoured products in brick-and-mortar stores, we identified a convenience sample of 18 vape shops across 15 different US states. All but one shop sold ice-fruit combination flavours. Thirteen shops sold ‘ice only’ flavours marked as containing only cooling features.

• Download figure
• Open in new tab
• Download powerpoint

Figure 1

E-cigarette solutions marked in ice-hybrid flavors: (A) Ice Dragon (blueberry, dragonfruit, mint). (B) Banana Ice (banana, menthol). (C) Drops Ice (menthol, mint). (D) Mixed Berry Ice (strawberry, blackberry, raspberry, blueberry, menthol), (E) Black (liquorish) Ice. See table 1 for additional description.

• Download figure
• Open in new tab
• Download powerpoint

Figure 2

Examples of disposable and pod-style e-cigarette products with ice-hybrid flavors: (A) Lush Ice (watermelon, menthol, raspberry). (B) Guava Ice (guava, menthol). (C) Lychee Ice (lychee, menthol). (D) Tropical Ice (pineapple, orange, mango, mint), (E) Iced Pineapple (JUUL compatible pod). See table 1 for additional description.

Our internet search results were consistent with Ramamurthi et al’s71 online search of disposable e-cigarette brand flavours on vendor websites, wholesale distributors, manufacturers and social media conducted June to August 2020. That study found that 33 of 139 total PuffBar brand flavours were marketed as fruit-ice hybrid flavours. For 449 different flavours marketed by non-PuffBar disposable e-cigarette brands, 34.5% were ice hybrid flavours, including fruit variants (eg, ‘Strawberry kiwi ice’), sweet drink variants (eg, ‘Cola ice’), and candy variants (eg, ‘Bubblegum Ice’).71 Ramamurthi et al also identified fruit-ice Puff Krush flavour enhancer caps that can be added to Puff Bar’s unflavored nicotine products and may circumvent certain regulations on flavoured e-cigarettes. They also found examples of disposable e-cigarette manufacturers who renamed some of their flavours from those that previously had ice themes to new non-ice concepts while retaining similar packaging (eg, ‘Icy Mango’ to ‘Marigold’). Hence, it is possible that ice-hybrid flavours may be more ubiquitous than these estimates.

Table 1 summarises manufacturer-provided and distributor-provided descriptions of ice-hybrid flavour products from our internet search. Descriptions of the ice-hybrid flavour products typically used terminology noting both fruity/dessert (eg, ‘sweet’, ‘citrusy’, ‘fruity’, ‘tropical’) and cooling (eg, ‘chill’, ‘icy cold’, ‘frosty’, ‘refreshed’, ‘shivering’) elements. The marketing imagery on the labels of ice-flavoured products often depicted ice cubes or snow alongside fruit or candy images (figures 1 and 2). Our internet search also found products with flavour names lacking the word ‘ice’ or a synonym but were nonetheless advertised to provide the combination of fruit/sweet and cooling (eg, ‘Apple’ by Naked 100, ‘Passion Fruit Orange Guava’ by Coastal Clouds Sweets mentioned on a 10 Fruit E-Liquids With Menthol Flavour for the Summer blog; not shown).72 Some ‘Ice only’ flavoured products marketed with descriptions of providing only the cooling sensations without any additional fruit or sweet elements (eg, Drops Ice; figure 1, table 1). Some distributor websites stated that the product contained menthol or used the term menthol to describe the quality of the cooling aspect of an ice-hybrid flavour (see table 1). Other distributors noted that the product contains ‘koolada’, which is a term loosely used to refer to various synthetic cooling agents.73 74

An internet search of message boards, consumer reviews and social media found user reports about ice-fruit combination flavours mirrored manufacturer-provided descriptors. One user described an ice-fruit e-cigarette solution as, ‘well balanced with a little berry flavour and iced finish…neither (flavour) is too overpowering’.75 Another ice-hybrid flavour product review stated, ‘great flavour, not too sweet, just right, and the iced part is spot on…leaves the sweet taste of berries and releases that cool sensation’.76

In sum, various ‘ice’ flavoured e-cigarette products are currently being sold that are marketed as having both fruity/sweet and cooling elements and are perceived by users as having both sensory attributes. Future systematic clinical experiments involving participants blinded to the flavour descriptors that collect sensory ratings are needed to confirm anecdotal reports and isolate that the cooling +sweet sensations in ice-hybrid flavours are caused by the products’ constituents and independent of how they are marketed. Additionally, research examining the effects of ice +fruit combination marketing themes (irrespective of the constituents in the product) on youth appeal may be warranted.

Non-menthol cooling agent formulations and self-mixing of ice flavors

As early as 2016, there are reports of retailers recommending their customers mix separately sold menthol-flavoured and fruit-flavoured e-cigarette solutions into a combination product.77 Our internet search found commercially available products identified as WS-23 ‘do it yourself’ concentrated formulations (eg, 30% WS-23 crystals dissolved in propylene glycol (by weight)). Manufacturers recommend that users add WS-23 concentrate to other fruit-only or dessert-only flavoured e-cigarette solutions.78–81 One WS-23 user review indicated that, ‘this is very potent, so use in very small amounts…even 0.5% in 15 mL was too much, but I can say without a doubt there is no menthol…in higher amounts it (gives a) fresh aftertaste but not menthol (taste)’.78 Another indicated, ‘this is awesome for making a nice cooling effect without changing the taste of your flavours. I mix at 0.8% for a mild ice effect. I don't like menthol, but I love this cooling effect with my fruit vapes.’. Another review referenced that the WS-23 cooling effect was stronger than menthol, stating, ‘Used 1/3 to 1/4 of what I had to use of my previous menthol stuff, and this was cooler’.

E-cigarette social media forums included discussions about variability in the sensory attributes of different cooling agents.82 One contributor posted, ‘WS-23…is the only real way I know of that you can get a super cold throat hit without any…flavour change’. Other postings from users indicated that WS-3 and WS-5 provide more ‘throat hit’ and ‘burn feeling’ than WS-23. One post stated, ‘WS-3…has a slight menthol taste to it and is less cool than WS-23’, whereas another post indicated, ‘WS-5 is stronger than WS-23 but can become bitter at higher percentages’. One exemplary comment noted that WS-23 ‘adds a “cooling” effect without the “taste” of mint or menthol’.

Tolerance was a recurring theme about non-menthol cooling agents in social media discussions.83 84 Some noted that increased frequency of vaping e-cigarette solutions containing WS-23 resulted in tolerance, which led users to increase the amount of WS-23 concentrate that they added to their e-cigarette solution in order to obtain the original desired effect. One person wrote, ‘Make certain you vape some uncooled juices every day, or you'll quickly need more and more cooling. It’s crazy easy to build a tolerance’.85 Another user indicated they developed a tolerance to high WS-23 concentrations and then temporarily abstained from using coolants. When they returned to using WS-23 again, somatic symptoms from the high WS-23 dose, stating ‘I laid off the coolant for a while and then tried (a high concentration of WS-23 again, which caused an) instant headache…Lowered it down…and it was much, much better. Tolerance is real’.83

Overall, some e-cigarette users prefer to self-mix e-liquids with synthetic coolant concentrates and may experience desensitisation or tolerance to the sensory attributes to aerosol with synthetic coolants. Combined with evidence about the pharmacology and toxicology of synthetic coolants described above, it is plausible that e-cigarettes with non-menthol coolants may cause certain physiological adaptations with high levels of use.