Friday, 20 November 2026

Erratum-watch: vaping risks

Patrick Gower's TV infotainment broadcast on vaping a few months ago referred to an NZMJ article claiming substantial potential cancer risk from vaping. 

That article included some claims that really didn't sound right.

For example, this one:

Klager et al. measured formaldehyde concentrations in vape smoke from four different vape liquid brands comprising 15 different flavours and reported concentrations in the range of 55.9mcg/m3 (45.6ppm) to 99,400mcg/m3 (81,011ppm).13 These concentrations all exceed the rat 14.3ppm exposure concentration, which led to increased nasopharyngeal cancer incidence and exceeded the lowest observable adverse effect concentration of 1ppm.7

The numbers of formaldehyde seemed really really high. Exposures 4 - 5,665 times the exposure causing cancer in rats seemed the kind of thing that would have been noticed. 

They similarly had this:

This variability is further evidenced by Margham et al., who reported formaldehyde/puff values between 17.5 and 59.0mcg.10 

Slightly lower than the reported Klager figures, but still quite high. Implausibly high. If the numbers really ranged from slightly over the amount needed to give cancer to rates to thousands of times higher than that amount, surely we would have heard it incessantly for years right? And regulation would look very different too.

The NZMJ published an erratum today. 

It has been brought to our attention that there are errors in our viewpoint article published in the New Zealand Medical Journal on 17 April 2026. In the section on formaldehyde, the concentration range of formaldehyde in “vape smoke” should be 17.5–59.0mcg/100 puffs, not 17.5–59.0mcg/puff, and later in the same section the concentration range for formaldehyde should be <0.0047ppm to 80.98ppm, not 45.6ppm to 81,011ppm. These errors do not change our overall conclusions that vaping represents an unquantifiable cancer risk and that our assessment supports the use of vaping for smoking cessation but does not exonerate vaping in its own right because taking up the habit introduces a new vaping-associated cancer risk.

So exposure was overstated by two to three orders of magnitude.

If you go back to Klager, I suspect something was just wrong with testing on the Vivid Vanilla flavouring. Is it more likely that that flavouring had formaldehyde at least 10x all the others, or that some weird overheating (or other mistake) affected that sample?




It is excellent that the authors found the error and published an erratum. 

But I wonder about the NZ Med Journal's refereeing process. Figures out by two to three orders of magnitude (100x and 1000x, respectively) is the kind of thing that ought to have stood out to a referee. 

If I'm now reading this right, the only observation in Klager on formaldehyde that's above the rat-cancer-risk threshold is that Vivid Vanilla sample that is suspiciously far higher than all of the other ones. 

No comments:

Post a Comment