An agenda for (useful) e-cigarette chemistry research

posted by Carl V Phillips

This is a bit off topic, but it follows from recent posts and this blog is where I have the most readership that will be interested.  Following up in particular on this post and the extensive comments therein (especially those from Mike Siegel, Konstantinos Farsalinos, and Spike Babaian) and in consultation with my colleague Igor Burstyn, the following is my assessment of what would be most useful for further research on the chemistry of e-cigarettes.

Right now, the research agenda seems to be driven by attempts by anti-harm-reduction activists to show that e-cigarette vapor contains measurable level of “toxins”, along with responses by others that in all the samples that have been studied, the levels of these toxins are below anything that we should worry about.  These have been done.  They are both established.  It is time to move on and do something that is useful for making sure that e-cigarettes are consistently as low-risk as is practical.

Researchers who have been studying cigarettes for decades have fallen into a laziness trap of believing they are in a static world where all exposures are exactly the same.  This is not true even in that world.  (Despite a favorite lie of the researchers in this area, it is obviously inevitable that some cigarettes pose less risk than others.  Indeed, the difference in risk among cigarettes probably dwarfs any differences among all the smoke-free alternatives, but that is a topic for later.)  The homogeneity assumption is even less true (when thinking in relative terms) in the world of e-cigarettes; variations from product-to-product and year-to-year are huge compared to the results of interest.

So the urge to do the simplest possible research, to just study a few e-cigarettes under one particular circumstance and imply that the results represent e-cigarettes in general, is misguided.  A study of a few particular products can only produce lasting knowledge by assuming all products are and will continue to be similar, which is not useful in general and is particularly not useful for understanding what might lead to a substantial increase in risk for a particular product.  At the low levels of contamination of interest, there will be too much variation for that extrapolation.

That said, there is no use in throwing out the data we already have, though to a remarkable extent that is what has been done because it is almost impossible to make sense of the existing studies.  With that in mind, the first item in the agenda for useful research would be to consolidate existing information from the several vapor studies that have already been done, from the original Ruyan study up to the present.  The results of these studies have been reported in such incommensurate ways, and so incompletely (with the exception of that first one, which provided the detailed reporting that is a defining characteristic for anything to deserve to be called “published”), that there may be a lot to learn from them by simply consolidating the information — particularly where the analyses discussed below are actually possible but simply were not reported.

This should include an effort to collect unreported information from the original researchers, which they should be willing to provide so long as they are interested in legitimate science and not cheap propaganda.  (Subtext:  It will be interesting to see if the US FDA provides their information, or if they are just going to admit that they fall into the latter category.)  It would be relatively inexpensive, and I am confident I could find someone who has not been involved in any of the studies or the debates to do it if the community and the industry could come up with funding — a small fraction of the funding that went into the recently published vapor study.

The main need for the research agenda, though, is creating results that offer generalizable information.  Future studies should focus on the mapping from e-cigarette liquid to the vapor, as a function of the technical specifics of the atomizer and e-cigarette hardware.  That is, instead of just learning what a few sample liquids and devices happen to create, in terms of vapor chemistry, what we really want to know is how to measure the liquid, consider the device, and then predict the vapor chemistry.

It will always be the case that measuring the liquid chemistry is relatively cheap and easy, while measuring the vapor is expensive and finicky.  Analyzing the liquid’s chemistry is cheap enough that it could become a standard part of the manufacturing process, or at least a frequently-used form of quality control.  But, of course, what we would really like to know is what is in the vapor that is produced from the liquid and inhaled, and so the mapping is far more useful than knowledge about one particular vapor sample.

Along the same lines, would be the mapping from vapor chemistry to the chemistry of what is exhaled by the vaper into the environment.  This is of rather less real practical interest, given how very little is actually exhaled and how minimal the apparent risk from vapor is even before this reduction.  But so long as there is political warring over “second hand vapor” there will be a demand for this also.  To the extent that anyone wants to study exhaled vapor, they should do it in ways that produce the most useful information, particularly comparing the vapor chemistry (sans human), but also looking at effects of behavior and comparisons to exhaled pure air (is that formaldehyde coming from the people or the vapor?)  Indeed, human subjects ethics demands that we get as much useful information as possible when we are using people in experiments.

Focusing future studies on the mapping will merely require doing the easier measurements alongside the difficult ones, and doing a decent job of reporting the full methodology and results (i.e., adhering to the standards of science, not of public health journals).  If someone is setting out to measure vapor chemistry, they just need to do the comparatively cheap and easy measurement of the same liquid’s chemistry at the same time, and also report exactly what was used to create the vapor.  Ideally this should include some useful technical measurements of the vaporization process, like what temperature was attained, but at least a report of the e-cigarette’s technical specifications would be of some use.

One of the biggest open questions about e-cigarette chemistry is whether some particular contaminants found in vapor are in the liquid already or are caused by the vaporization process, and if the latter, is this is an inevitable result of the entire technology or are there ways to reduce it (should there be any health concern at all).  It is really quite amazing that in 2012 we still only have what might be called “decent guesses” about this.  This relates to the more general question of whether it is enough to know the liquid chemistry to be confident that the unintended unhealthful exposures are trivial, or do we need to know something else.

There is no excuse for wasting resources doing e-cigarette chemistry studies that do not seek to determine the mapping.  There is almost no downside, other than a modest increase in the cost of the study.  The effect on information produced will be entirely positive; the usual throw-away result — what happened to be in this particular sample from this particular product, on particular day — will still be there if someone really thinks there is value in it.  Indeed, that will be doubled because there will be information about the chemistry at two stages.  But the real value will be the relationship between them.

[UPDATE: This conversation continues here.]