Sounding today more like an episode of Midsomer Murders than a title stemming from a classic of children’s literature, the Bears of Little Brains is a reminder that those of us who work in science can all to easily become bears of little brains, falling into our own heffalump traps, and setting ourselves on the trail of a woozle. We may publish something only to find, being a Bear of Very Little Brain, that a Thing which seemed very Thingish inside us is quite different when it gets out into the open, and has other people looking at it. And so it was with the second Walach publication, the research letter about children wearing masks inhaling high concentrations of carbon dioxide, which Dr No mentioned in a comment to his last post. The demolition jobs done on twitter of Walachs et al’s research letter were swift and vitriolic, and the comments tab on the letter itself is turning into an ever lengthening dossier of the delusions held by the letter’s authors. Retraction is increasingly looking inevitable.
Dr No will put his hand up straight away, and say he was guilty of succumbing to confirmation bias. He saw a publication that suited his purposes, and promptly made use of it. He has no excuse for not being more careful, given that only days before (on 27th June 2021) he rejected another paper that also had Walach as the lead author. Walach the Wally has form, and form tends to persist. In his defence, Dr No pleads that he only described the research letter as ‘interesting’ — the main problem for Dr No is the methodology uses a proxy, rather than direct measurement of arterial carbon dioxide levels — and the link only appeared in a comment, rather than getting a full post of its own. But he did use the letter to say ‘another good reason to get schools back to normal now‘, and that does imply he thought the findings were credible.
Over the last few days, the research letter, and the responses it has generated, have continued to niggle Dr No. The letter is flawed, for sure, but some of the comments it has generated suggest the technology exists to do the study properly, so the really interesting question becomes why has no one done the proper study? Dr No may have missed it, but neither his own searches, nor the references supplied in the many and varied comments, lead to a straightforward well conducted study reporting arterial carbon dioxide levels in healthy children wearing face masks. A ‘mini review‘ published in February 2021, covering publications up to 7th November 2020, concluded there were no usable studies on the question. The one study the review found that might have been useful did find raised, but not clinically significant, PETCO2 levels (more on them shortly), but bombed because it only covered short term (five to ten minutes) of mask use, and cannot be generalised to long term use.
Since the mini-review, there appears to have been just one potentially usable study, but it too is flawed. The research was done in May and June 2020, but for some reason publication was delayed until March 2021. Median PETCO2 levels were reported, and remained remarkably constant, almost too constant, in the older (but still young) group, but the groups were small (22 for age 0 to 24 months, 25 for age 2 to 12 years), and again the study only covered short term, in this case 30 minutes, of mask use. It is also unclear how the subjects were recruited, with the report perhaps suggesting they may have been loitering about an Italian secondary-level hospital paediatric unit. Perhaps they do things differently in Italy. Like the previous study, this study cannot be generalised to long term use, and furthermore, it cannot be generalised to older children. We remain over 18 months into the pandemic, and still no one has done the proper study.
Slowly the reasons started to appear. Dr No has already pointed out in the comments to his last post that the truly proper study has major ethical problems, in that it requires repeated invasive and painful blood tests be done on healthy children, to measure actual arterial carbon dioxide levels. The ethics of doing research that involves painful procedures on healthy children can be complex, all the more so when it is a third party, the parent, who gives or withholds consent, but generally ethical decisions rightly go against conducting such research. Indeed, Dr No has form on this, having declined to get involved in a highly dubious study that would have involved taking blood samples from healthy kids to detect almost certainly non-existent anaemia.
The ethical objections to invasive procedures means using a proxy, or surrogate measure, gained using a non-invasive method, which is what Walach et al and the other two studies linked to above attempted to do. The deficiencies in the former have been noted in the comments to the letter; the other two studies used something called capnography, which measures PETCO2 levels, to ring their bells. All these long words are very troubling for bears of little brains, who can all to easily get their campanologies and capnographies mixed up. Capnography uses (near) real time measurement of carbon dioxide concentration in the breath right at the very end of exhalation (the ET in PETCO2 stands for end tidal) to get a better estimate of arterial blood carbon dioxide levels. The general idea, that the last of your gasp contains the closest you will get to undiluted alveolar air from deep inside your lungs, and so will reflect blood, albeit venous, carbon dioxide levels, sounds plausible, but in practice it too is flawed.
Capnography is typically used in acutely ill patients in high dependency settings like ICUs. In these settings, it has a known bias, or tendency to under-estimate arterial carbon dioxide, of about 7mmHg, which can be taken into account when interpreting results. There do not appear to be any data on whether a similar, greater or lesser bias applies in healthy kids. Nor does capnography adapt particularly well to monitoring free range kids, because of the equipment involved. Even if these hurdles are overcome, there is a major problem: the PETCO2 is dependent on the length of exhalation, which in turn depends on respiratory rate. This happens because exhaled air carbon dioxide concentration increases during exhalation, as the breath gets closer and closer to undiluted alveolar air from deep in lungs. You need a long and full exhalation to get a reading that reflects alveolar carbon dioxide levels; short and shallow breathing on the other hand cuts short the exhalation phase, so the end tidal carbon dioxide level does not have time to reach alveolar levels, resulting in significant under-estimations of arterial carbon dioxide levels.
This matters, because kids wearing masks typically increase their respiratory rate (see Walach’s raw data here, and Table 2 here, from the second linked study above), meaning PETCO2 will almost certainly under-estimate arterial carbon dioxide levels. Not only that, but there is another intriguing, but unanswered, question: is the increase in respiratory rate a physiological response to raised arterial carbon dioxide levels? Capnography may be a notch above Walach’s suck it and see approach, but it is still far from perfect. The PETCO2 levels in the two limited capnographic studies were within normal limits, but were they valid measures of arterial carbon dioxide concentrations? Unfortunately, the short answer is we don’t know. Even if more studies were done in more naturalistic settings for more realistic periods of mask use, we still wouldn’t know whether the data under-estimates arterial carbon dioxide levels, and if so, by how much.
There is yet another non-invasive way of estimating arterial carbon dioxide levels, using transcutaneous measurements. This method is more convenient, needing just a sensor attached to the skin, and it generally works well for steady state conditions, such as sitting in class for hours on end. Dr No cannot find it, but there must be some reason why this method has not been used to estimate arterial carbon dioxide levels, because no one has done it. The bottom line is that there is as yet no real world, ethically acceptable, valid and reliable method, with the possible exception of transcutaneous monitoring, but no one has used that, of measuring arterial carbon dioxide concentrations in healthy children wearing face masks in natural settings.
We are back to Dr No’s old favourite, the distractions of getting exact answers to the wrong question, set against appreciating good enough answers to the right question. On this occasion, we don’t even have a good enough answer to the right question. Dr No suspects, but cannot prove, that we are being distracted by exact answers to the wrong questions, and, if that is not enough to make him continue to call for a return to full normality in schools, then let us not forget that raised arterial carbon dioxide levels is only one concern of many about ongoing face masking for school children. As the image for this post shows, masks damage social interaction, and for kids, this is happening at a time when they are learning to interact socially. In the absence of (a) solid proof of significant covid–19 transmission in schools and (b) equally solid proof that masks do not raise arterial carbon dioxide to harmful levels in normal children, then there can be no justification at all for inflicting the monstrosities of masking on normal children. Even this bear of little brain can see what that means: masks off now!