Not only were the average subtest scores no different in patients treated with HBO and NBO, the average performance of patients who received NBO was normal or above normal in every domain. Recalling that the original primary outcome was DNS defined solely by neuropsychological testing, the figure makes clear that the study would have been resoundingly negative had the original outcome been retained.

In short, despite their positive framing, the studies of both Thom and Weaver could, for different reasons, be legitimately interpreted as negative. But even when viewed as positive, they do not provide compelling evidence of benefit from HBO.

Where does this leave us? It should not lead us to conclude that HBO has no role in the management of CO poisoning. There are clearly grounds to speculate that it might. If so, the effect is likely to be small.(2) The available data do not tell us which patients are most likely to benefit, how much real benefit they might reasonably expect from treatment, what the optimal HBO regimen might be, or whether the benefits of therapy justify the potential harms and costs, since treatment sometimes entails the transport of patients over long distances to a hyperbaric chamber.

Every day, thousands of patients around the world are poisoned by carbon monoxide. Some die in the prehospital setting, but many more present in need of treatment. Clinicians who elect to treat them with HBO do so in the hope that it might help, not because they know it will—operating, as Seger puts it, on nothing more than informed guesswork.(19) This same phenomenon plagues much of clinical toxicology, but in this instance the problem is hardly an insoluble one. There is ample justification for a large, multicentre RCT of HBO in CO poisoning to answer, once and for all, the important questions that remain unanswered.

Such a trial is not only justified, it is exigent. Yet we face a formidable barrier to its implementation—one that was articulated more than 15 years ago(9) and must be articulated yet again. For such a trial to come to fruition, we must help those who hold the keys to the hyperbaric chambers understand that the evidence for HBO in CO poisoning is not as compelling as they would believe.

If, another 15 years from now, we find ourselves no further ahead on this issue, it will represent not only a missed opportunity to properly evaluate a potentially useful treatment relevant to countless patients around the world each year, it will constitute nothing less than a monumental failure of the field of hyperbaric medicine.