Toxicity classification methodology
Toxicity classification for all chemicals reported by Vardavas et al. was sought in the classification and labelling information database of the European Chemicals Agency [13]. The EU provides clear guidance on the estimation of toxicity of chemicals and mixtures through the CLP regulation [10, 11]. Each chemical is classified according to different hazards. There are mainly 3 types of hazard classes: physical hazards, health hazards and environmental hazards [14]. Based on the type of hazard and toxicity classification, specific hazard statements are required each of which has a specific code [15].
Health hazards include acute toxicity for oral, dermal and inhalation exposure, with chemicals being allocated to one of four toxicity categories (Category 1 to 4) according to specific numeric criteria. For these health hazards, acute toxicity values are expressed as approximate LD50 values or as acute toxicity estimates (ATE) from experimental data [10]. The classification categories are defined according to dose cut-off values of chemicals (in mg/kg body weight) causing toxicity in animals, with higher dose needed to cause toxicity corresponding to lower toxicity classification. E-cigarette liquids are mixtures, with flavouring chemicals diluted in non-toxic solvents (PG and VG). Therefore, the method of classification of mixtures for toxicity was used to identify the toxicity classification of each chemical at the concentrations reported by Vardavas et al. Since test data on the mixture itself or similar mixtures are not available, the classification was based on calculation thresholds. Two types of analyses were performed. In one, each compound reported in the study by Vardavas et al. [8] was assumed to be the only component of the mixture dissolved in non-toxic solvents (PG and VG) at the maximum levels reported by Vardavas et al. A limitation of this method is that electronic cigarette liquids contain more than one flavouring chemical that could have a toxicity classification. Unfortunately, the previous study did not provide information on the composition of each of the liquids tested. Thus, and to address this limitation, we calculated the toxicity estimate for a hypothetical final solution (e-cigarette liquid) containing all the flavouring chemicals at the maximum concentration reported by Vardavas et al. This is performed using the additivity formula (10), which involves adding the Acute Toxicity Estimate (ATE) of each ingredient for each hazard classification, in this case of all compounds at the maximum concentrations reported by Vardavas et al. While it is unlikely that an e-cigarette liquid would contain all these flavouring chemicals at the maximum reported concentrations, it provides an estimate of the worst-possible case scenario based on the study findings. Thus, to estimate the hazard classification for mixtures, we used the following formula:
$$ \frac{100}{ATEmix}=\frac{\sum \limits_n Ci}{ATEi} $$
where ATEmix is the acute toxicity estimate of the mixture containing a specific concentration of the chemical, n is the number of ingredients (one ingredient for the analysis of each compound separately and sum of all ingredients in the analysis of a liquid containing all ingredients in maximum concentrations), Ci is the concentration of the chemical i in the mixture, and ATEi is the converted acute toxicity point estimate of chemical i. Since toxicity classification to different categories is based on a range of acute toxicity estimates, we used the converted acute toxicity point estimates as recommended by the CLP regulation in the calculations [10]. It should be clarified that lower ATEi represents lower exposure dose needed to cause toxicity and thus higher toxicity classification (i.e. more toxic). Similarly, lower ATEmix represents lower exposure dose of the mixture needed to cause toxicity and thus higher toxicity.
For other health and for environmental hazards, such as skin corrosion and irritation, eye irritation, respiratory irritation and toxicity to aquatic life, percent concentrations of the chemical are used to determine the different toxicity categories (Table 2). For these hazards, the maximum concentrations reported by Vardavas et al. were used. Separate analyses were performed considering that each chemical represents a unique mixture with the chemical present at the maximum concentration and assuming that an e-cigarette liquid contains all ingredients at the maximum concentrations reported. In the latter case, the concentrations of each compound having a specific toxicity classification were added in order to calculate the toxicity classification for the final mixture.