Artificial intelligence (AI) has proved a key enabler in recent research from RCSI, which revealed the potentially harmful substances produced when e-liquids in vaping devices are heated for inhalation. The study, published in Scientific Reports, underscores the urgent need for public health policies concerning flavoured vapes.
The research team at RCSI’s Department of Chemistry leveraged AI to simulate the effects of heating e-liquid flavour chemicals found in nicotine vapes. By including all 180 known e-liquid flavour chemicals, the AI predicted the new compounds formed when these substances are heated within a vaping device immediately prior to inhalation.
The AI analysis revealed the formation of numerous hazardous chemicals, including 127 classified as ‘acute toxic’, 153 as ‘health hazards’, and 225 as ‘irritants’. Notably, these included volatile carbonyls (VCs), which are known to pose significant health risks. The AI predicted the most popular fruit, candy, and dessert-flavoured products as primary sources of these VCs.
Lead author Professor Donal O’Shea, Professor of Chemistry and Head of Department, expressed concern over the findings: “We wanted to understand, before it’s too late, the likely impact flavoured vapes are having on the health of the growing number of vapers. Our findings indicate a significantly different profile of chemical hazards compared what we are familiar with from traditional tobacco smoking.”
The research suggests we may see a new wave of chronic diseases linked to current vaping practices manifest over the next 15 to 20 years. By highlighting these findings, the research team aims to empower individuals to make more informed health choices and contribute to a broader dialogue about the long-term risks of vaping and the need for comprehensive regulatory approaches
The study also highlighted the complexities introduced by the vast array of flavours available in vaping products, which include 180 different chemicals blended in various amounts. This cocktail of chemicals, primarily derived from the food industry where they have a good safety record for specific uses, were never intended to be heated to high temperatures for inhalation. The AI framework established in this study could lead to the development of risk reports for individual flavours, providing a valuable public health policy resource.
Considering the popularity of flavoured vapes among non-smoking teenagers and young adults, understanding the long-term effects of these products on public health is crucial. This study demonstrates that without comprehensive regulation, there is a substantial risk of transferring new health issues to younger generations.
The research was conducted in collaboration with IBM Research – Tokyo and supported by Taighde Éireann – Research Ireland through the Pathway Programme for Dr. Dan Wu.
You can read more in the paper ‘Forecasting vaping health risks through neural network model prediction of flavour pyrolysis reactions’ in Scientific Reports.
