Initial screening of e-liquids was performed on a JEOL JMS T100LC Accu-ToF DART-MS using a previously published method (Poklis et al., 2015). In brief, a capillary tube was dipped directly into the e-liquid and then introduced into the helium stream for analysis to identify components’ exact mass. The data was analyzed by creating an averaged, background subtracted, centroided mass spectra that was calibrated using PEG 600. Data was evaluated using National Institute of Standards and Technology (NIST) and Scientific Working Group for the Analysis of Seized Drugs (SWGDRUG) libraries. Full method details can be found in Supplementary Material A.

E-liquids were also screened using a simple dilute-and-shoot preparation with an untargeted analysis method employed on a Shimadzu QP-2020 GC-MS (Kyoto, Japan). Samples were processed using NIST 17-1, 17-2, and 17s libraries and SWGDRUG 3.5-3.9 libraries for identification. Full method details can be found in Supplementary Material A.

Quantitation of nicotine was accomplished by GC-MS using a Shimadzu QP-2020 GC-MS following previously published parameters (Pagano et al., 2016). Full method details can be found in Supplementary Material A.

Quantitation of caffeine and menthol was accomplished by developing a single-ion-monitoring method for GC-MS using a Shimadzu QP-2020 GC-MS. Full method details can be found in Supplementary Material A.

Quantitation of VE and VEA was accomplished by developing a single-ion-monitoring method for GC-MS using a Shimadzu QP-2020 GC-MS. Full method details can be found in Supplementary Material A.

Quantitation of acetone, ethanol, isopropanol, and methanol was accomplished using a modified version of a previously published method for headspace gas chromatography-flame ionization detector (HS-GC-FID) (Poklis et al., 2017b) and employed a Shimadzu HS-20 headspace sampler attached to a Nexis 2030 GC-dual FID (Shimadzu Corp., Kyoto, Japan). Full method details can be found in Supplementary Material A.

Quantitation of olivetol was accomplished by developing a multiple-reaction-monitoring method for LC-MS/MS using a Shimadzu LC-MS 8050. Full method details can be found in Supplementary Material A.

Quantitation of cannabinoids was accomplished using a modified version of a previously published method with a Shimadzu LC-MS 8050 (Poklis et al., 2010). Full method details can be found in Supplementary Material A.

Propylene glycol and vegetable glycerin were the most commonly identified carriers in e-liquids. Both are considered GRAS and used in food products, pharmaceuticals, and health and beauty products. Though considered harmless based on years of use in food and medicine, use of these chemicals in e-cigs and vaporizers is not inherently safe. Studies have demonstrated that when heated to high temperatures, like those of a heated e-cig coil, PG and VG can produce carbonyls such as formaldehyde and acetaldehyde (Kosmider et al., 2014; Geiss et al., 2016; Troutt and DiDonato, 2017; Qu et al., 2019; Jiang et al., 2020). Device settings have been found to directly affect the production of these harmful chemicals, and therefore may lead to the risks of increased exposures to these carcinogens, as carbonyl formation has been shown to increase directly with increasing battery output voltages (Kosmider et al., 2014; Qu et al., 2019).

PG and VG were the only known carriers identified in nicotine and flavorant-only formulations. In addition to PG and VG, MCTs, PEG, and squalene were identified in cannabinoid and DOTN formulations, as well as formulations that contained both nicotine and a DOTN. These carriers may be used to dissolve more lipophilic compounds, including cannabinoids, before adding them to PG and VG. MCTs may also be used because they are attributed with health benefits. MCTs have been touted online as a safer, healthier alternative to PG and VG (Zachar, 2018). MCTs produce harmful volatile organic compounds and increase interleukin-8 (IL-8) and interleukin-6 (IL-6) levels which are biomarkers for lung inflammation and injury. MCTs also decrease transepithelial electrical resistance and increase lipid-laden macrophage formation which can lead to lipoid pneumonia (Muthumalage et al., 2020b). MCT aerosols were found to contain alkyl alcohols, which are surfactant-like and can produce cytotoxic effects. MCTs and PEG both produce harmful carbonyls when aerosolized (Troutt and DiDonato, 2017; Jiang et al., 2020). PEG was found to produce levels of formaldehyde that neared those seen by traditional combustion cigarettes, with one puff exposing the user to 1.12% of the daily exposure limit set forth by the United States Occupational and Safety Health Administration (OSHA) (Troutt and DiDonato, 2017). Squalene, a clear, slightly yellow oily substance, has been implicated in causing exogenous lipid pneumonia (Cha et al., 2019).

Recent studies have documented the presence of both VE and VEA in THC products (Taylor et al., 2019; Muthumalage et al., 2020a; Duffy et al., 2020). VEA is thought to be used as a diluent that mimics the consistency of a high purity THC oil, thus deceiving the consumer into believing they are purchasing a high-grade THC product and allowing a larger profit for the manufacturer by extending the supply (Duffy et al., 2020). VEA is also implicated as one of the primary chemicals thought to be responsible for the EVALI epidemic (Blount et al., 2019; Blount et al., 2020; Muthumalage et al., 2020b). Though VEA was not identified in any samples, VE was identified in multiple cannabinoid-containing products. Both VE and VEA produce quinone-like compounds, which can produce reactive oxygen species that increase cytotoxicity (Jiang et al., 2020).

Trends in the use of chemicals used as flavorants are difficult to discern due to limited sample size (241 products) compared to products available commercially, the variety of chemicals available, the constantly changing formulations, and evolving product regulations. Though most flavorants have been designated as GRAS, the majority of the flavorants identified through screening are associated with at least one GHS classification, such as irritant, corrosive, or acutely toxic, which contraindicates their safety.

Several studies have demonstrated that flavorants produce varying degrees of toxicity to cells through different mechanisms (Behar et al., 2016; Leigh et al., 2016; Sherwood and Boitano, 2016; Behar et al., 2018). Omaiye et al. reported total concentrations of flavor chemicals in refill fluids were found as high as 362.3 mg/ml, while pod and cartomizer formulations evaluated (JUUL and Vuse, respectively) only contained total flavor chemical concentrations of 0.2–15.7 mg/ml (Omaiye et al., 2019a). They also reported that all 8 evaluated JUUL liquid formulations and the corresponding aerosols were cytotoxic. Cinnamaldehyde is already widely reported as cytotoxic, and a study evaluating cinnamaldehyde in e-liquids found that increasing battery output voltage further increased cytotoxicity (Behar et al., 2016). Costigan et al. pointed out that use of the same e-cigarette liquid in different devices will alter the aerosol formation (Costigan and Meredith, 2015). These studies’ findings suggest that a e-liquid toxicity may vary based on the device and operational settings.

Another flavorant-related health concern identified in this study is the formation of flavorant-carrier adduct products. Compounds such as vanillin propylene glycol acetal are formed post-production, sometimes within hours of initial product mixing, through an acetalization reaction between an aldehyde and PG. The formation of these acetal products was reported to be dependent on the ratio of PG present in the formulation (Erythropel et al., 2019). Adducts were shown to aerosolize with similar efficiency as the parent aldehyde and proved more effective at activating respiratory irritant receptors than the respective parent aldehyde, suggesting e-cig liquids may become more harmful to the user as time passes and flavorant-PG adducts form. The Flavor Extract Manufacturers Association (FEMA) lists some flavorant-carrier adducts as their own unique entity, which can be purchased for use as a flavorant (Flavor and Extracts Manufacturers Association of the United States, 2021). This can further complicate the determination of product ingredients in terms of what is added by the manufacturer or what was formed post-production. Additionally, if these chemicals are intentionally added, their concentration within the products may rise above the manufacturers intended level due to the post-production formation.

The combinations of flavorants may potentiate the harmful or toxic effects of one another. Manufacturers are currently not compelled to release e-liquid formulation recipes nor list all ingredients used within products on labels. Relevant toxicity studies are impossible to conduct as the proprietary formulations rapidly evolve.

Nicotine, while a common constituent, is not always present. Some e-cig liquids are manufactured as nicotine-free, flavor-only options, and sometimes other pharmacologically active ingredients are used, as exemplified by the numerous cannabinoid liquid options available today. In this study, one e-liquid contained nicotine even though the product label claimed it did not contain any amount of nicotine. The presence of nicotine in this product could be due to contamination during the manufacturing process, incomplete/insufficient labeling, or the intention of including a known pharmacologically active ingredient to elucidate some effect. Other products were found to have higher or lower concentrations of nicotine than were labeled on the product, again demonstrating poor quality assurance and quality control standards in the industry (Peace et al., 2016; Peace et al., 2018).

Nicotine concentrations increased as manufacturers switched to nicotine salt formulations, which use nicotine with an organic acid, such as benzoic or lactic acid. These formulations allow manufacturers to significantly increase the nicotine concentration while reportedly reducing the harshness of such high nicotine amounts (Harvanko et al., 2020). The highest levels observed in this study were from JUUL liquid formulations, which use nicotine salts. The concentrations observed in these products correlate with concentrations identified in other studies (Omaiye et al., 2019a).

While nicotine use and dependence are well documented in the scientific literature, toxicity and poisonings, especially as it pertains to the e-cigarette industry, are worth discussing. Higher nicotine concentrations increase the chance of accidental nicotine poisoning, both through inhalation or ingestion of the liquid. The American Association of Poison Control Centers has reported thousands of poisoning cases about e-liquids since 2011, when e-cig use became more prevalent in the Unites States. As of May 31st of this year, 2063 cases have been reported (National Poison Data System, American Association of Poison Control Centers, 2021). Many poisoning cases involve young children ingesting the products accidentally, while some cases involve someone intentionally ingesting or injecting the liquid for a means of self-harm. Two cases of poisoning are reported following inhalation of a nicotine product by active duty military personnel, leading to clinical nicotine toxicity requiring emergency medical services (Bendel et al., 2021). Dermal contact from spilled or leaky pods is a concern that should not be overlooked, as nicotine readily absorbs through the skin, leading to both localized and systemic health concerns. Attempted homicide by nicotine liquid being poured directly onto the skin of the victim has been reported, with the victim describing the liquid as sticky with a spicy flavor (FOX 9, 2021).

Caffeine was identified and quantitated in JUUL Menthol and Classic Tobacco liquid formulations. JUUL Classic Tobacco products contained an average of 23.5 μg/ml caffeine, while JUUL Menthol products contained an average of 9.3 μg/ml. In another study, caffeine in JUUL Menthol and Classic Tobacco aerosols were found in concentrations of 0.037 and 0.090 mM, respectively, showing it is able to both aerosolize and be inhaled by users (Omaiye et al., 2019a). Caffeine was not labeled on the e-liquid products.

Caffeine affects the cardiovascular, renal, nervous, and respiratory systems (Ueno et al., 2020), and is widely consumed throughout the world as a legal stimulant. In addition to the caffeine consumed in coffees, teas, sodas, and energy drinks, it can be found as a dietary supplement and is used in narcolepsy and asthma therapies.

Issues regarding caffeine in e-cig liquids should be considered. Inhalation of caffeine increases its bioavailability. An in vivo study using mice demonstrated caffeine inhalation, via a nebulizer, was an effective way of administering caffeine and produced greater spontaneous activity compared to the same dose administered intraperitoneally (Ueno et al., 2020). Additionally, caffeine was identified as an unlisted ingredient. Individuals with caffeine sensitivity or underlying medical conditions that require a caffeine-free lifestyle could be endangered by inhaling caffeine.

The addition of caffeine to e-cig liquids could act as an initiation primer, leading to increased caffeine seeking and consumption and chances of caffeine addiction. Caffeine consumption has been reported to increase the odds of smoking, the urge to smoke, and the subjective reinforcement from smoking (Treloar et al., 2014). A correlation between combined inhalation of caffeine and smoking with promotion of coronary heart disease and severe vascular lesions has also been reported (Pan et al., 2021).

Menthol was identified in a variety of products evaluated, sometimes as a listed ingredient, sometimes unlisted. Menthol was quantitated in concentrations as high as 4.48 mg/ml. Another study reported finding menthol in concentrations up to 68 mg/ml (Omaiye et al., 2019b). In that study, the cytotoxic properties of several flavorants were evaluated. Menthol was identified as toxic in cells studied in concentrations 30 times lower than their highest identified concentration, meaning all products evaluated with menthol in concentrations greater than approximately 2.5 mg/ml would be considered cytotoxic.

Menthol has been identified as having other important pharmacological properties related to smoking and vaping. Therefore, its identification as an additive in non-menthol flavored products is not surprising. Menthol is reported as imparting a cooling sensation with analgesic or counterirritant effects, reducing the perceived harshness of the nicotine and smoke or aerosol (Ton et al., 2015; Taylor et al., 2021). Menthol’s effects are thought to allow users to inhale deeper, hold the smoke or aerosol in the lungs longer, and use products with higher nicotine content, all which may allow toxic or carcinogenic chemicals to have a longer duration of exposure (Ton et al., 2015). Menthol, while itself is an irritant, can also work as a counterirritant for other chemicals commonly encountered in e-cig liquids.

Menthol has been reported to reduce nicotine metabolism both in-vitro and in-vivo, leading to increased systemic nicotine exposure, reduced clearance, and longer durations of pharmacological effects (Benowitz et al., 2004; Alsharari et al., 2015). Menthol has been associated with reduced Cytochrome-P450 2A6 isoform activity (MacDougall et al., 2003; Benowitz et al., 2004). This same enzyme is responsible for nicotine metabolism, thus co-ingestion of menthol with nicotine will alter nicotine metabolism and elimination and therefore prolong the pharmacological effect of nicotine felt by the consumer.

Ethanol has been identified in many of the products evaluated in this study, including nicotine refill formulations, nicotine pods, and DOTN formulations, with concentrations ranging from not-present up to 217.2 mg/ml. Previous studies have also reported on ethanol concentrations in e-cig liquids, and have also identified concentrations of ethanol greater than 20% (Valentine et al., 2016; Poklis et al., 2017b).

Some flavoring chemicals use ethanol as part of the manufacturing process, such as vanilla, which is required by FDA Code of Federal Regulations to contain “not less than 35% by volume” (United States Food and Drug Administration, 2020a). While ethanol may be present in formulations due to their use in flavorings, it can also be added to thin the liquid, to help dissolve other substances that are not miscible with typical carriers, or for intentional consumption of ethanol. Do-it-yourself (DIY) e-liquids may contain higher concentrations of ethanol than those found in the manufactured products reported in the literature.

There is a dearth of knowledge regarding ethanol pharmacokinetics and intoxication from inhalation. Inhalation bypasses first-pass metabolism and increases bioavailability of ethanol. It has also been suggested that co-administration of ethanol with nicotine could lead to increased dependence and addiction liability for both substances due to their synergistic nature (McKee et al., 2006; Oliver et al., 2013), especially as ethanol has been found to potentiate several of the rewarding effects of nicotine (Rose et al., 2004).

A clinical study from 2017 compared the effects of inhaling a nicotine e-liquid with high or trace amounts of alcohol (Valentine et al., 2016). They found no difference in subjective effects between high and trace alcohol groups. Their findings suggest that users of high alcohol concentration e-cig liquids may experience some alteration in psychomotor function without recognizing any subjective effects to alert them to the impairment. Additionally, they found the metabolites of ethanol in the urine of three participants out of the eight exposed to the high alcohol e-liquid.

Olivetol was identified and quantitated in five cannabinoid-based e-liquid products. Concentrations ranged from below the limit of quantitation to 3.9 mg/ml. Olivetol is a naturally occurring organic compound that can be used as a precursor for synthesizing various cannabinoids (Tadayon and Ramazani, 2021). Commonly found in lichen, it also exists for a short time in cannabis plants before conversion to CBGA, the precursor to THCA and CBDA. Its presence in e-liquid products may be due to an incomplete chemical reaction used by manufacturers when synthesizing THC, and thus could be considered an impurity. Olivetol may also be an intentional ingredient in e-liquid products. It is touted as an “antidote” to purportedly reduce unwanted effects from the consumption of high concentration THC products, such as anxiety, paranoia, or feeling overly “high” from a THC overload (Carberry, 2018; Royal Queen Seeds, 2019). Olivetol is thought to act on cannabinoid CB1 and/or CB2 receptors (Carberry, 2018). Evidence of olivetol’s ability to reduce some effects of THC was reported in the UNDOO, LLC product patent application (Carberry, 2018). No clinical trials evaluating olivetol’s effects have been conducted to support these claims. UNDOO reports volunteer testimonies from real-life, non-clinical trials in which participants either smoked or ingested THC, then took known amounts of olivetol, and reported on their subjective experiences.

Olivetol is listed in the National Institute of Health’s PubChem database as a GHS irritant, causing skin, eye, mucous membrane, and respiratory irritation that could lead to severe tissue destruction and specific target organ toxicity with a single exposure ( National Institute of Health, 2021 ). It emits carbon monoxide and carbon dioxide when heated to decomposition, and it is recommended to use a NIOSH-approved respirator equipped with an organic vapor/acid gas cartridge when handling neat olivetol (National Oceanic and Atmospheric Administration, 2021).

Cannabinoids have been identified and quantitated in several products evaluated in this study. CBD and THC were the most abundant in both prevalence and concentration and were quantitated in concentrations as high as 332.6 and 497.7 mg/ml, respectively. Cannabinoid-based e-liquid formulations have existed since the advent of the modern e-cigarette. JUUL is a spin-off company and product of PAX, a discreet cannabis vaporizer launched in 2012. Grenco Science also officially launched the THC vaporizer in 2012, after years of product development and testing (Freedman, 2014; Farah, 2017; Bobrow, 2019; Hartman, 2021). The launch of these cannabis-based e-cigs coincided with adult-use legalization in Colorado and Washington. Even though almost every state in the United States has legalized some form of C. sativa, whether medical or adult use, the regulation of cannabis and cannabis products vary by the state.

The National Institute on Drug Abuse (NIDA) Monitoring the Future survey data indicates that while “any vaping”, “vaping nicotine”, and “vaping flavors” trends appear to be steady, or maybe even slightly decreasing, between 2019 and 2020, “vaping marijuana” is still on the rise among 8th, 10th, and 12th graders (National Institute of Health, 2020).

Like other chemicals, cannabinoid labeling accuracy has significant deficiencies. In this study, products were analyzed that were labeled with the wrong concentration, or listed as THC-free, though THC was identified. Supplementary Material D lists bottle/packaging claims and results of screening and cannabinoid quantitation. Eight out of nine products analyzed that indicated total volume of product or cannabinoid content were over-labeled. By comparison, one study reported nearly half of the CBD products analyzed were under-labeled and about 25% were over-labeled, reporting “vaporizing liquids” to be the most frequently mislabeled (85%) and oil the most frequently labeled accurately (Bonn-Miller et al., 2017). Additionally, they reported THC was present in 18 of 84 products tested, though they were listed as THC free. The current study found THC in five samples out of nine that indicated they were “THC free” or contained “0% THC”. Inaccurate product labeling demonstrates the lack of quality assurance and quality control required in the industry and poses a significant danger to consumers who may consume a higher dose than intended.

Chronic use of products containing trace concentrations of THC can result in failed urine drug tests (Spindle et al., 2020; Sholler et al., 2021). Additionally, higher peak blood concentrations have been reported from vaping THC compared to the same dose of smoked cannabis (Spindle et al., 2019). This increase in delivery efficiency is thought to be a product of minimized sidestream smoke and lack of drug pyrolysis, both of which reduce the possible dose to be inhaled in traditional combustion delivery methods (Pomahacova et al., 2009). Compared to combustion smoking conditions, vaping cannabis has been found to increase the frequency of testing above immunoassay cutoff levels in a clinical setting (Spindle et al., 2020).

Some marketed “cannabis” products contain synthetic novel psychoactive substances (NPS) as the active drug. Many NPSs are not scheduled, and therefore legal, at the time of product manufacturing. By the time these chemicals are identified and federally scheduled, manufacturers have adopted another NPS which is not scheduled, allowing manufacturers to skirt federal DEA regulations. These NPS can be more potent and can lead to severe and life-threatening situations. Some consumers have information that certain products contain NPSs and knowingly choose to use those products. Some consumers are unaware that products purchased for relief contain NPSs but experience untoward effects (Poklis et al., 2019).

Cannabinoid-based products evaluated in this study have generally become more complicated over the years. Early formulations were mainly comprised of PG, VG, cannabinoids, and terpenes. More recent formulations contain a variety of carriers, sometimes mixing multiple carriers in one product, as well as extra active ingredients and flavorants. Terpenes found in older products are thought to be carried over from extraction methods (Peace et al., 2016). Recent formulations market terpene-specific profiles to appeal to flavor preferences or purported health benefits. Until these products are federally regulated, formulations can only be limited to individual States’ regulations.

The carriers, diluents, thickeners, flavorants, and solvents identified have been generally considered as safe because of their accepted safety for oral ingestion, yet there is minimal to no evidence for long-term consequences from inhaling these substances. These chemicals can create injury to the lung tissue. They can prevent proper oxygen flow, disrupt cell membranes, cause irritation and inflammation to the lung tissue, mucosa, and bronchi, and induce lipoid pneumonia ( Erythropel et al., 2019; Thirion-Romero et al., 2019; Muthumalage et al., 2020b ). Two mechanisms for lung pathogenesis have been proposed. The first hypothesis describes acute exposure that creates a direct chemical injury that results in negative health effects (Alexander et al., 2020). The second mechanism describes a change to the immune cells in the alveoli due to chronic exposure to a chemical, which may or may not result in symptomology recognized by the vaper. With the addition of a new chemical, the body reaches some threshold that triggers a pathologic inflammatory response, precipitating neutrophil recruitment, edema, and necrosis (Johnson and Matthay, 2010; Alexander et al., 2020).