Search

Flavour ingredients are an essential part of e-liquids. Their responsible selection and inclusion levels in e-liquids must be guided by toxicological principles. We propose an approach to the screening and toxicological risk assessment of flavour ingredients for e-liquids. The screening involves purity requirements and avoiding ingredients that are carcinogenic, mutagenic or toxic to reproduction. Additionally, owing to the uncertainties involved in potency determination and the derivation of a tolerable level for respiratory sensitisation, we propose excluding respiratory sensitisers. After screening, toxicological data on the ingredients should be reviewed. Inhalation-specific toxicological issues, for which no reliable safe levels can currently be derived, can lead to further ingredient exclusions. We discuss the use of toxicological thresholds of concern for flavours that lack inhalation data suitable for quantitative risk assessment. Higher toxicological thresholds of concern are suggested for flavour ingredients (170 or 980μg/day) than for contaminant assessment (1.5μg/day). Analytical detection limits for measurements of potential reaction and thermal breakdown products in vaping aerosol, should be informed by the contaminant threshold. This principle leads us to recommend 5ng/puff as an appropriate limit of detection for untargeted aerosol measurements.

Mathematical models have been developed for the toxicological risk assessment of chemical mixtures. However, exposure data as well as single chemical toxicological data are required for these models. When addressing this data need, it could be attractive to focus on chemicals with similar mechanisms of action, similar modes of action or with common target organs. In the European Union, efforts are currently being made to subgroup chemicals according to this need. However, it remains to be determined whether this is the best strategy to obtain data for risk assessment. In conditions such as cancer or HIV, it is generally recognised that pharmacological combination therapy targeting different mechanisms of action is more effective than a strategy where only one mechanism is targeted. Moreover, in diseases such as acute myocardial infarction and congestive heart failure, several organ systems concomitantly contribute to the pathophysiology, suggesting that a grouping based on common target organs may also be inefficient. A better option may be to prioritise chemicals on the basis of potency and risk of exposure. In conclusion, there are arguments to suggest that we should concomitantly consider all targets that a chemical can affect in the human body and not merely a subset.

The aim of this paper is to show that the uncertainty factors used in toxic risk assessment to develop exposure standards do contain societal judgments as well as technical judgments. The process generally used today originated in the 1950s, when a deterministic approach to risk was the norm. Technical judgments are required concerning the nature and the quality of the evidence used in the risk assessment. Judgments taken are essentially cautious. This caution may not matter when measured exposure is significantly below the standard and may be accepted when exposure occurs only following an approval process based on “gate keeping.” More sophisticated judgments are required when actual exposure may exceed this type of standard or when risk needs to be compared with benefit. These circumstances can occur with patient exposure to human medicines and with occupational exposure to chemicals. Under these circumstances more explicitly considered societal judgments concerning what constitute “broadly acceptable” and “tolerable” risk criteria, and hence what are appropriate uncertainty factors, are required. The outcomes of those societal judgments are likely to vary according to the circumstances surrounding the exposure and have led to smaller uncertainty factors being considered appropriate for occupational exposure, when compared with widespread public exposure.

In recent years, there has been increasing awareness of the differences in maximum allowable levels developed by the various federal regulatory agencies. It has generally been presumed that the variation in levels arises from differences in risk management decisions, i.e., how the data are used. This study was undertaken to test the hypothesis that differences in choice of data to utilize also have a significant impact on interagency variation. To test this hypothesis, a comparison was made between the data bases used by the EPA Office of Drinking Water and the Occupational Safety and Health Administration in setting maximum levels for pesticides in drinking water and workplace air, respectively. The results show that the same data are only infrequently used as bases for both types of regulatory levels. The study also indicates that the differences in data selection cannot be ascribed to factors related to the applicability of different data for the two types of exposures—drinking water and workplace air. There are essentially no differences in the data with respect to route of exposure, species (laboratory animals vs humans), or length of exposure. No obvious scientific basis for the differences was found. Other possible selection factors are discussed.