There's an emerging theory that the profound hypoxia by COVID is due to a sort of hemoglobinopathy caused by direct viral protein binding to heme. I suppose anything is possible, but I'm highly highly doubtful and feel we should address this head on. 1/10

#1 -- Hbopathies are pretty easy to identify in the ICU. They're not infrequent. Pulse ox measures the saturation of Hb with oxygen, but we also do routine arterial blood gases, which delivers the PaO2, a partial pressure of oxygen in plasma. 2/10

Lung's main duty is to get O2 to plasma. There, it (usually) rapidly binds to Hb. Lungs don't work, PaO2 is low and Hb sat/pulse ox low. Hbopathy, PaO2 fine and pulse ox low. We have tools for figuring things out when this happens -- CO-oximetry, metHb, Hb electrophoresis. 3/10

Seen no reports of PaO2/pulse ox discrepancies. #2 -- infiltrates are some sort of direct damage by heme/Fe. Don't know how heme would be released without hemolytic anemia. Seen no evidence. This doesn't fit clinical pictures of iron overload in lungs -- hemosiderosis or 4/10

hemochromatosis are more chronic and restrictive. #3 -- Some point to high compliance being not typical of ARDS. Not true, large amounts of research devoted to optimal vent strategies in heterogeneous disease, which very well may have high compliance. This is a main fear 5/10

for balancing baro- and atelectotrauma in ARDS. ARDS is a clinical umbrella that encompasses many disease processes, just defined by hypoxia + noncardiogenic pulmonary edema. There could absolutely be microthrombi or dysregulation of hypoxic pulm arteriole vasoconstriction. 6/10

Certainly worthy of continued investigation and aggressive empiricism but can't forget what we know to work -- ARDSnet ventilatory strategies and supportive care. #4 -- the data that this theory is based on is a preprint of in silico work. I know something about in silico 7/10

protein prediction and it's next to useless without wet work. Clinical decisions absolutely absolutely can not be made on this alone, at least give us some evidence that this can happen in a dish, preferably in an animal or a human! #5 -- This is somehow given as further 8/10

evidence for use of hydroxychloroquine/vitamin C whatever. That there's some resistance to using whatever regimen is being touted today. Nobody is saying not to use them -- they're saying we need to collect data to know if they work. 9/10

By this point, I'd imagine many are on these drugs from early stage of disease anyway. What's clear is that they are not miracle cures, thus careful data collection is needed to see what effect they are having (if any). 10/end