A Systematic Failure

Seeing as one of the leading tobacco control ex-purts tweeted about this particular analysis, I decided to have a look. Well after he tweeted this, how could I refuse?


Well, seeing as I’m not an industry stooge – as the “Caped Crusader” would like to believe – I’m not actually going to “attack” it. Not as such anyway. Seeing as he asked soooo nicely.

So, what’s this analysis all about then? Well, as you expect from Gloomy Creepwell it’s all about the “risks” from e-cigarettes. Specifically, the “risks” from passive exposure – i.e. “second-hand vapour” – to e-cigarette vapour. I did wonder when they’d get around to trying to drum up some more fear and confusion about this. Y’see, second-hand smoke is a myth and second-hand vapour is no different. Well, OK it is slightly different in that there is a distinctive smell to vapour as opposed to tobacco smoke. But that’s about it.

Right off the bat let’s make one thing absolutely crystal clear. This is a systematic review, could almost call it a meta analysis, but that would be unkind to those. In effect, this looks at x number of actual studies (which will have varying methodology and objectives) to try and say that A+B+C+D=Z, when in fact A+B+C+D will probably equal something completely different. It really doesn’t matter how many regression analysis, or any other fancy number crunching you do, if the input is crap the output will be crap. The shit-in, shit-out principle.

Does this mean that reviews are meaningless? Not really, they do have some value just not a substantial amount – it might be better if the review (in this case) looked at x number of studies with similar (though not necessarily the same) methodology, but that really is neither here nor there – end of the day, the PHE report was nothing more than a glorified systematic review which reached a controversial conclusion in its 95% figure. Was that report meaningless? Not in the slightest, it was a comprehensive review looking at a much broader array of factors. This review however, only looks at one specific angle. Is it invalid? Probably not, but that does depend heavily on the studies that the authors “reviewed” – so let’s take a look shall we?

As with any review, the authors searched through a series of journals – in this case Embase, Ovid Medline and PreMedline – for the time-frame 1996 to 10 September 2015. Which immediately strikes me as odd. There have been a few studies published since September 2015 which would fit the search terms (e-cigarette/s, e cigarette/s, electronic nicotine delivery, vaping, vape, vaper/s – combined with ‘passive’, ‘secondhand’, ‘second hand’ or ‘exposure’, ‘exposed’, or ‘vulnerable’ or ‘nonuser/s’ or ‘non-user/s’) – either way, the search terms identified 462 studies. Out of that 462, 150 were duplicates leaving only 312 for relevance to e-cigarettes (and ‘health effects’) excluding titles describing regulation, perceptions, advertisements and uptake studies (137). The remaining 175 abstracts were then reviewed.

So from 462, the authors were left with 175 to read the abstract. Most abstracts don’t go into a whole lot of detail (which is why it is called an abstract) so simply reviewing the abstract may have generated false negatives – i.e. studies that were subsequently excluded because the abstract didn’t contain the information the authors were looking for:

  • Looking at health effects from passive exposure to e-cig vapour (animal or human)
  • Testing or analysing e-cig vapour directly (like that daft cell study) or
  • Testing or analysing ambient air with vapour (much like those sneaky gits at the Vape Convention)

Studies were excluded if they looked at:

  • Health effects of the user only (direct exposure)
  • In utero exposure
  • Third hand exposure (that apparently is a thing now)
  • Ingestion or dermal exposure to e-liquid

Based on those requirements, the 175 studies were then whittled down to 24 for a “full text review”, and a further 8 were excluded at that point, leaving only 16.

Out of the final 16 the studies (which included one conference abstract – seriously?) were then grouped into four general study designs:

  • Direct passive exposure – human volunteers (4)
  • Direct passive exposure – animal models (1)
  • Indirect exposure – human volunteers (with or without a smoking machine) – (7)
  • Indirect exposure – no human volunteers (4)

A quick scan of the studies (helpfully listed in Table 1) suggests that many of the studies had extensive limitations – small sample sizes, unclear on applicability to humans, indirect study only, no baseline reading, cross-contamination, limited measurements and so on – the list is pretty extensive, which begs the question – why were these studies even selected? But of course, we already know the answer to that don’t we?

For instance, how would you even consider that “third hand exposure” (The living-room air was sampled for nicotine for 7 days, and saliva and urine samples were collected from the volunteers after this week of exposure. The results showed significantly higher levels of airborne nicotine in homes with EC users than in nonsmoking control homes. In homes with CC smokers, airborne nicotine was significantly higher than in homes with EC users.) having any relevance to “potential health risks”?

In short, you can’t. In fact, that particular study suggests that while nicotine is indeed present in both a smokers home and a vapers home, there was significantly more found in a smokers home – not really a problem is it?

In another experiment, Flouris et al. exposed 15 nonsmokers for 1 hour to second-hand CC smoke or EC vapour generated by a smoking machine. Serum cotinine and lung function measures were taken for each participant. No difference was found in lung function for the nonsmokers passively exposed to EC vapour compared with no exposure, but participants’ serum cotinine levels were raised, similar to volunteers passively exposed to CC smoke.

So expose folk to combustible cigarette smoke (what other kind is there?) take measurements, let them return to baseline then expose them to vapour. Find similar levels of the nicotine marker – cotinine – in both cases, but crucially, lung function was not affected. So no health risk there either.

The review goes on and on and on about the sixteen chosen studies, which when taken as a standalone paper don’t really tell anyone much – the two examples above show just how slim the straws the authors are grasping really are. The key to a review such as this is always going to be in the conclusion. Which is cringeworthy.

Studies examining the composition of EC vapour or some of its aspects found that ECs are not emission-free.

Well no, they aren’t “emission free” – but then again, define “emission” and at what level. Will that level have any kind of biological effect? Toxicology 101 people. Dose. Poison. That kind of thing.

The majority of studies concluded that passive exposure to EC vapour may pose a health risk to bystanders.

Actually, they didn’t. It’s that biological effect thing again. There was mention of “possible risk” but not actually concluding that vapour may pose a risk. Different wording, completely different meaning. Context.

It is noted that those studies undertaken by tobacco employees or funded by the National Vapers Club concluded no apparent risk from ECs to bystanders. Those who did not declare a conflict of interest were more likely to draw conclusions that were more precautionary and/or suggested a potential risk from passive exposure to ECs, highlighting potential biases in the current literature.

I’ll just say snork here.

When examining the absolute impact from passive exposure to EC vapour – that is, comparing EC vapour with background levels – EC vapour contains elevated levels of nicotine, Particulate Matter, glycerine, propylene glycol, formaldehyde and acetaldehyde, PAHs and metals. These studies demonstrate that EC vapours can contain harmful chemicals and have an impact on indoor air quality.

Of course vapour contains nicotine you dolts. I’ll just link to Dr Farsalinos for the nicotine, glycerine and propylene glycol comment, and Clive Bates for the Particulate Matter. The fact that something (in this case “harmful chemicals” – don’t get me started on chemicals, everything is a chemical ffs) is detected does not mean it’ll have any kind of biological effect.

So what do these folks say in conclusion?

Well, you guessed it:

Although more research is required, current evidence regarding passive exposure to EC vapours shows the potential for health impacts. Those passively exposed to the vapours of EC users are exposed to numerous pollutants at levels above background and at concentrations that are associated with potential adverse health effects. The risk from being passively exposed to EC vapour is likely to be less than the risk from passive exposure to CC smoke.

In a nutshell, these things have the potential for harm, more research is needed, oh and we just “don’t know”.

Sigh – out of 462 studies, this is the best they can come up with?

D- : must try harder.

  • Chris Oakley

    I am not an industry stooge either. I occasionally write things that lead to such accusations partly because I despair at the continued acceptance of garbage such as the review you highlight and the ongoing abuse of science by people pushing authoritarian political agendas. Having to pay for people such as Capewell simply adds insult. to injury. My tolerance and respect for people who work in public health does not extend to extremists such as Capewell and Martin McKee. Their behaviour is unacceptable and their continued tenure is a stain on the reputation of UK academia.

    • It’s something a little special isn’t it when ordinary everyday folk like you and me are classified as “industry stooges”. Which is why I have this blog, so I can take the mick 😉