E-cigarette flavors and aldehyde emissions: another failure to verify findings from a previous study



By Dr Farsalinos

More than 1 year ago, a study was published in the journal Environmental Science and Technology reporting that aldehyde emissions from e-cigarettes are derived from thermal degradation of flavoring compounds rather than the main liquid ingredients (propylene glycol and glycerol). The study found non-detectable levels of aldehydes in unflavored liquids, but up to 10,000-fold higher levels of aldehydes in flavored samples.

At that time, I uploaded a comment clearly stating that the results contradict previous research on aldehyde emissions. I also emphasized that the findings were NOT related to overheating and dry puffs, unlike other studies that had seriously overheated the devices. In my comment I was very careful not to mention anything insulting for the authors of the study, since I had no proof that something went wrong and I could not find any flaws in the publication. I did however mention that: “One of the fascinating aspects of science is the ability to replicate, and thus confirm or reject, the findings of a study… We also plan to use some of the liquids used in the recent study (same brand), if the author replies to my email requesting information on the brand used (there is no such information in the manuscript, and I still have not received any response to the email I have sent).” I also remind everyone that I submitted a letter to the editor of the journal presenting my views on the issue, obviously before the replication attempt was made.

The main author of the study responded to my comment in what I consider a clearly provocative and insulting tone.  He mentioned: “Dr. Farsalinos revealed that (a) he is not up-to-date with the current literature, and (b) has not read our paper carefully”. I suggest everyone to read my previous comment and the comment by the main author Dr Khlystov in order to understand the difference in the context.

As I had promised, we were planning to replicate the study. After 2 emails I sent to the main author asking to report the brand of the liquids tested (the e-cigarette device and power settings were mentioned in the manuscript, but not the liquid brand), and after many public requests by myself and Prof Peter Hajek through PubMed Commons (unfortunately all PubMed Commons comments have been removed from PubMed), the authors NEVER reported the liquid brand they tested. In PubMed Commons, the main authors referred us to the cigalike brand they tested, which was useless because we could not use unflavored liquid in the prefilled cartomizers of the cigalike and thus could not compare flavored with unflavored liquids. Additionally, prefilled cartomizers are known for their inconsistent performance, making such comparisons difficult. In my opinion, this was an unprofessional behavior. Additionally, the main author NEVER responded to any of my emails requesting for such information. I have never experienced in the past such a behavior, and unfortunately I must say that this was unprofessional.

We were able to replicate the study due to a coincidence. One of the flavors used in the study was called “Dragon’s Café”. We could find only one company (a US company) which produced a flavored e-cigarette liquid with this name. We also found all other flavors from this company as they were tested in the original study, and chose the 3 flavors with the highest levels of aldehydes (Dragon’s Café, watermelon, blueberry). The liquids were available in standard and sweetened versions. Considering that sweeteners can be transformed to aldehydes when exposed to heat, we chose both the standard and most sweetened versions. The study was replicated using the same devices and power settings as the original study, while additionally we tested similar flavorings from a different manufacturer and a newer generation e-cigarette device.

For start, no dry puffs were detected with any of the liquids. This verified my original assessment as documented in my first comment and the letter to the editor. Our results were revealing in identifying a very small contribution of flavorings on aldehyde emissions. In fact, aldehyde levels were so low that consumption of 5 grams liquid per day would expose vapers to less formaldehyde and acetaldehyde than just staying at home and breathing air. For acrolein, exposure was orders of magnitude lower compared to NIOSH-defined recommended safety limits. To give you an idea of the differences in results, the authors of the original study found up to 7000 ug/g formaldehyde, while we found a maximum of 62 ug/g.

This is the third replication from our group which failed to verify previous findings. I should note that  the journal Environmental Science and Technology has been involved in two of these replications, publishing 3 papers (paper 1, paper 2 and paper 3). The replication rejecting the findings of papers 2 and 3 can be seen in a previous comment.

In a recent review on aldehyde emissions from e-cigarettes, I stressed the need to replicate studies reporting unusually high levels of aldehyde emissions. For example, I reported that only 4 of the 32 published studies verified that no dry puff conditions were generated during the laboratory experiments. It seems that in this case there are more methodological issues, possibly related to false-positive results. The authors of the flavorings study are now presenting findings that e-cigarette use increases aldehyde levels in exhaled breath. Considering their previous publication on flavorings and aldehyde levels, I will be very reluctant to accept their results unless they are independently replicated. If they publish the results, we will try to perform a replication study. There is already a publication showing that the exhaled breath of vapers has the same level of aldehydes as non-smokers. So, everyone understands my concerns about another study reporting “peculiar” results. However, I am not going to rush into judgment, I will wait for their results and the replication (if we will be provided with details on the study materials).




Background Image

Header Color


Content Color