Secondhand vaping: the (lack of) evidence

A deep dive into the evidence on passive vaping

Last month, the British government launched a consultation on whether to ban vaping in all workplaces, including bars, clubs, pubs, shops and restaurants. This is a solution in search of a problem. There has been no groundswell of support for such a ban. Even ‘public health’ groups have not been lobbying for it. The government’s Impact Assessment (IA) says it will cost more than £500 million but is not able to put any monetary value on the benefits because it is unable to establish if they will be any benefits.

Looking for a rationale for banning vaping in six million buildings, the Impact Assessment raises the issue of whether ‘secondhand vapour’ may pose a risk to the health of bystanders. It cites several studies and reports looking at this question. Let’s take a look at them.

The WHO

The first is from the World Health Organisation, an agency that is extremely hostile to vaping and that is constantly receiving community notes on X (Twitter) for spreading misinformation about it. The IA says:

“the WHO state that the vapours generated by vapes typically raise the concentration of particulate matter in indoor environments and contain nicotine and other potentially toxic substances. Vape vapours therefore pose potential risks to both users and non-users”.

Note the use of the word “quantifiable” rather than “harmful”. Even the tiniest quantities are quantifiable with modern technology. No environment - inside or outside - is entirely free of particles in the air that could be harmful in large enough amounts.

The WHO citation is just a webpage in the format of Frequently Asked Questions. Under ‘Are secondhand ENDS [electronic nicotine delivery systems - the pointless ‘public health’ term of e-cigarettes] emissions dangerous?’, it says:

The aerosols generated by ENDS typically raise the concentration of particulate matter in indoor environments and contain nicotine and other potentially toxic substances. ENDS emissions therefore pose potential risks to both users and non-users.

The IA repeats most of this verbatim, but the WHO does not cite any supporting evidence. Note that even the WHO does not go beyond talking about “potential” risks.

The EU

The second source is an EU council recommendation on “smoke- and aerosol-free environments” from 2024. It leans heavily on the WHO and, like the WHO, it supports a ban. Most of its citations for secondhand exposure involve heated tobacco, but it does cite five peer-reviewed studies about e-cigarette vapour which, it says, show that e-cigarette emissions “expose bystanders to quantifiable levels of particulate matter and key toxicants and contaminants”. The IA quotes this verbatim, but note, again, the use of the word “quantifiable” rather than “harmful”.

The five studies are:

Fernández et al. (2015) found that “PM2.5 [particles that are 2.5 micrometres or smaller in diameter] median concentration was 9.88 μg/m3 in the e-cigarette user home and 9.53 and 9.36 μg/m3 in the smoke-free homes”. The authors admit that “the PM2.5 medians in the e-cigarette user home and non-smokers smoke-free homes were similar” and “whether secondhand exposure from e-cigarettes poses health risks at short- and long-term is still unknown.” To put those figures in context, the concentrations were 572.52 μg/m3 in a smoker’s home.

Li et al. (2020) is a literature review from some California-based researchers. They show their hand early on when they note the rise of vaping in China and other countries and predict that “secondhand exposures to e-cig aerosols will likely become a potential public health problem in those countries soon”. They say that research into the health effects of secondhand vaping is “inconsistent”, but report unusually large readings of PM2.5 in various exceptional or unrealistic settings, such as at a vape conference, in an exposure chamber with a smoking machine and in a lab where the measuring device was positioned 50cm away from the vaper.

Hess et al. (2016) is another literature review, this time by Australian researchers. They found very little evidence of harm. For example:

“[The study] concluded that there is no significant risk of harm to human health from exposure to the levels of tested chemicals.”

“No difference was found in lung function for the nonsmokers passively exposed to EC [e-cigarette] vapour compared with no exposure, but participants’ serum cotinine levels were raised” [cotinine is a biomarker for nicotine and is often used a proxy for exposure to nicotine]

“This study found that short-term passive EC exposure did not seem to lead to the inflammatory response that is seen in volunteers passively exposed to CC [conventional cigarette] smoke – blood count measures were unchanged”

“the short-term exposures to EC vapour did not elicit a reduction in lung function or an increase in inflammatory markers.”

“As found by Flouris et al., this short-term passive exposure to EC vapour did not significantly affect inflammatory markers in the exposed subjects.”

They nevertheless conclude that “current evidence regarding passive exposure to EC vapours shows the potential for health impacts.”

Lerner et al. (2015) subjected mice to “acute exposure of e-cig aerosols” and found that this led to “measurable oxidative and inflammatory responses in lung cells and tissues that could lead to unrealized health consequences”. Studies involving humans have not found the same effect, probably because humans are not exposed to unrealistically high doses.

Borgini et al. (2021) was a quasi-natural experiment in a room with two habitual vapers. Its main finding was that Juul e-cigs produce less aerosol than a brand called Just Fog. The researchers didn’t even bother looking at PM2.5 because there was so little of it. (Roberto Sussman has written about this study in more detail.)

The EU document also cites a study by Glantz et al. (2024) which has nothing to do with secondhand emissions. I don’t know why they referenced that. Perhaps because Stanton Glantz is a fanatical opponent of vaping?

Studies specifically cited in the Impact Assessment

The IA mentions three other studies. They are:

Amalia et al. (2023) went looking for chemicals and aerosols in the homes of 29 vapers. They didn’t find nicotine in eight of them and in the rest the average level was “quite low”. That’s putting it mildly. The average level was 0.01 μg/m3 . That’s one hundredth of a microgram per cubic metre of air. In studies of secondhand smoke, this would be considered below the limit of quantification. There was no statistically significant difference between levels of particulate matter in vapers’ homes and non-vapers’ homes. Non-vapers living in houses where people vaped had higher levels of cotinine in their salvia and urine, although vastly less than vapers themselves. The authors called for further research. (Roberto Sussman has discussed this study in more detail.)

Tattan-Birch et al. (2024) found that children living with vapers absorb more cotinine than children who don’t. But, as the data in the study show, the quantities were homeopathically small. Cotinine in the blood of secondhand vapers averaged 0.08 μg/L (micrograms per litre) whereas secondhand smokers averaged 0.49 μg/L, and children exposed to neither vape nor smoke averaged 0.02 μg/L. To put that in context, the average vaper has a blood cotinine reading of around 250 μg/L, more than 3,000 times higher than that seen among those who are “exposed” to secondhand vapour. (NB. cotinine is harmless.)

Islam et al. (2022) All research on vaping from California comes with a red flag and this is no exception. This cross-sectional study used data from two waves of the a health survey and suffered from massive problems of confounding because, of the 223 young adults who reported being exposed to secondhand vapour, only 6.9% were not exposed to secondhand vapour, tobacco smoke or cannabis smoke and did not smoke or vape themselves. 11.5% of them were vapers and/or smokers but supposedly had no exposure to second-hand smoke. 48.4% were smokers and/or vapers and were also exposed to secondhand vapour/smoke and 33.2% didn’t vape or smoke and were exposed to secondhand smoke, but not to vaping. Teasing out the effects of secondhand vapour from this muddy dataset seems an impossible job. The researchers attempted to do so and found a 40% increase in bronchitis and 53% increase in shortness of breath among the secondhand vapers. Experts were critical of this study when it was first published, not least because the effect size is implausibly high (it suggested that passive vaping is worse for you than actual vaping).

Studies cited by OHID

Finally, the IA cites the extensive report on vaping from the Office for Health Improvement and Disparities (OHID), published in 2022. This, in turn, cites the following six studies:

Martínez-Sánchez et al. (2019) got six people who didn’t vape or smoke to live with a vaper for a week. After seven days, their urine was tested for tobacco-specific nitrosamines (NNAL). Two of them didn’t have a detectable level of NNAL. The other four averaged 0.47 pg/mg (picograms per millilitre; a picogram is one trillionth of a gram). The authors correctly note that this is “very low” and that the average level among smokers is “around 300 pg/mL”, i.e. 638 times higher.

Johnson et al. (2019) sent 34 people to an e-cigarette convention. If you’ve ever been to one you’ll know that there is nothing vapier than a vape convention. After spending six hours amongst cloud-chasers, the participants had their urine and saliva tested for acrolein, tobacco-specific nitrosamines and cotinine. Acrolein was found at average levels ranging from 163 to 839 μg/g. The authors note that there are “many sources of acrolein exposures both in the environment and endogenously, thus the acrolein concentrations measured in this study are only partially attributable to passive e-cigarette exposure.” Levels ranged between 186 to 324 μg/g before the participants attended the event.

As in other studies, passive vaping was associated with higher levels of cotinine, peaking at 2.31 μg/g of urine after one event, but generally being below 1 μg/g. These levels are not unusual among non-smokers, including those who are not exposed to tobacco smoke. When researchers use cotinine tests to identify smokers, they use a threshold of 100 μg/g or more.

In contrast to Martínez-Sánchez et al. (2019) (see above), the researchers found that all tobacco-specific nitrosamines tested for (including NNAL) were “below the limit of detection in all samples for all sampling times and sampling events”.

Qunitana et al. (2019) is some more research from California. One of the authors of this paper is Georg Matt who invented the concept of ‘thirdhand smoke’ which, like ‘passive vaping’, relies on scaring people about barely measurable levels of chemicals in their environment. These researchers gave some kids wristbands which measure nicotine and then tested their urine after a week. They found a close correlation between wristband readings and urine readings. That’s it. No data on how levels changed relative to exposure. All it did was show that the wristbands work.

It’s not obvious what this study contributes to the literature although the authors do point out that the wristbands are commercially available, and one of the authors declares an interest in the company that makes them.

Melstrom et al. (2018) In this study, six people who don’t use nicotine “were exposed to secondhand aerosol from ad libitum e-cigarette use by three e-cigarette users for 2h during two separate sessions” and then had their blood, saliva and urine tested. Suspiciously high baseline readings suggested that at least one of them was a secret smoker. There was an increase in average cotinine levels after the experiment but only very slightly, e.g. blood readings went from 0.299 ng/ml to 0.326 ng/ml and saliva went from 0.350 ng/ml to 0.395 ng/ml. Is that a lot? It’s less than the variation among the participants before the experiment began when it ranged between undetectable and 0.894 ng/ml for blood and undetectable to 1.056 ng/ml for saliva. The authors, who include the future head of the FDA’s tobacco division, nevertheless conclude that their findings “could have important implications for public health policy” because “aerosols produced from the use of e-cigarettes may contain other harmful constituents that were not measured in this study.” This is typical of the way researchers in this field sideline their own findings in an attempt to justify vaping bans.

Qunitana et al. (2021) It’s those Californian wristband advocates again. This time they produce some half-useful data. Nearly all the kids’ wristbands reported some exposure to nicotine, but kids who lived with vapers had eight times more of it than kids who didn’t and kids who lives with smokers had ten times more of it more than the kids who lived with vapers. The silicone wristband had now been adapted to measure tobacco-specific nitrosamines but only five out of the 26 wristbands on kids who lived with vapers showed any detectable level of these. The conclusion is once again pro-wristband: “Silicone WB [wristbands] show promise for sensitive detection of exposure to tobacco-related contaminants from traditional and electronic cigarettes”.

Amalia et al. (2021) put a vaper in a car and a room and measured the air for particulate matter (PM2.5). They also measured nicotine, cotinine, 3’-OH-cotinine, nornicotine, tobacco-specific nitrosamines, propylene glycol, and glycerol in the saliva of “bystanders”. The highest median PM2.5 reading was 21 μg/m3 in the room and 16 μg/m3 in the car, and “most of concentrations of the airborne nicotine and all biomarkers were below the limit of quantification in both settings.” A few bystanders reported “mild irritation symptoms”.

Although PM2.5 levels roughly doubled after the vaping began, levels of PM2.5 recorded outdoors before the experiment began were within the same range (14-25 μg/m3). The authors admit that “the highest median PM2.5 concentration in our study did not exceed the outdoor guidance level of World Health Organization air quality standard (25 μg/m3 as daily average) and the United States Environmental Protection Agency Air quality index (35 μg/m3 as daily average)”, but insist that the levels were “not negligible”. That is a matter of opinion. They were less than you’d getting from lighting a candle and much less than you’d get from cooking food.

Moreover, there is a fundamental difference between PM2.5 from vapour and PM2.5 from burnt matter. As Roberto Sussman points out, “particulates of exhaled e-cigarette are highly volatile liquid droplets almost entirely composed of glycerol, propylene glycol, nicotine and water” whereas “PM2.5 particles from atmospheric air pollution exhibit a very complex chemical compositions, largely consisting of semi-volatile and non-volatile compounds, mostly generated by combustion sources: primary and secondary organic carbon, a wide variety of sulphates and nitrates, hydrocarbons, metals and crustal material”. The misleading conflation of vape and smoke (or, more precisely, of liquid droplets and particulate matter), with the implication that they are both equally harmful, is common in these studies.

Conclusion

That is all the evidence the UK government has put forward, directly or indirectly, on the ‘risks’ of secondhand vapour. Most of the studies are from researchers who are actively looking for risks and who write up their work in a way that emphasises the “potential” harm. They generally fail to provide adequate context by referring to typical readings among active vapers, let alone active smokers, and they rarely refer to the safe thresholds of the substances they are examining (the 2021 study by Amalia et al is one of very few exceptions). That is because the levels recorded are generally considered safe by regulators in workplaces and outdoors (which is where the regulations tend to be applied). Unable to show that the measurements are unsafe or abnormal, the researchers focus instead on an increase in one substance - usually cotinine - and imply that any increase above the baseline must be hazardous.

It should be noted that these are the studies mentioned by organisations such as the WHO who want vaping banned indoors. Weak as it is, they presumably think that it is the best evidence to support their position, but other evidence is available. For example, when the U.S. Department of Health and Human Services sampled the air in a vape shop (where e-cigarette use was obviously heavy in a confined space) it found that all chemicals in the air were below the occupational exposure limit. It expressed concerns about detectable levels of two chemicals (diacetyl and 2,3-pentanedione), but both of these are banned for use in e-cigarette fluids in the UK (and EU).

Similar studies have found that even in very high exposure conditions in a small, non-ventilated vape shop, nicotine concentrations in the air were undetectable and those chemicals that were detectable were at very low (and legal) levels.

A systematic review of the evidence found “no evidence of potential for exposures of e-cigarette users to contaminants that are associated with risk to health at a level that would warrant attention if it were an involuntary workplace exposures” and “no evidence that vaping produces inhalable exposures to contaminants of the aerosol that would warrant health concerns by the standards that are used to ensure safety of workplaces.” And that is to the users of e-cigarettes! “Exposures of bystanders are likely to be orders of magnitude less, and thus pose no apparent concern.”

Public Health England said in 2016 that “there is no evidence of harm to bystanders from exposure to e-cigarette vapour and the risks to their health are likely to be extremely low.” They also said that “e-cigarette use is not covered by smokefree legislation and should not routinely be included in the requirements of an organisation’s smokefree policy”. Why? Because there is no risk to bystanders and vaping bans discourage smoking cessation.

As Prof Peter Hajek, Director of the Tobacco Dependence Research Unit at Queen Mary University of London (QMUL) says:

“While health risks of e-cigarettes to vapers themselves have been estimated at up to 5% of health risks of smoking, health risks to bystanders are most likely reduced by a much bigger margin, and most likely altogether. This is because e-cigarettes release no chemicals into the environment themselves, only what users exhale, and such exhalation has so far not been shown to generate any toxicants at levels that could conceivably affect the health of bystanders.”

There has been a concerted effort by anti-vaping academics to find evidence that ‘secondhand vapour’ is harmful to bystanders. Despite using a variety of methods, they have come up empty-handed, with the partial exception of a few studies that have looked at air quality in unrealistic laboratory conditions. The levels of chemicals measured in the atmosphere and in the bodies of people ‘exposed’ to vaping in everyday situations are not only vastly lower when compared to tobacco smoke, but are lower when compared to everyday activities such as cooking and are consistently below the safe level for indoor and outdoor air quality.

Update

If you want an even deeper dive, Dr Sussmann has followed up this post here and here.