Granted, the boundary between organic disease psychosomatic / functional neurological disorder might be demarcated by "is it detectable by well-validated clinical-grade diagnostics (in-vitro or otherwise) or not"; which I think is pure map/territory confusion. Under this definition, long covid is indeed psychosomatic; there is no such diagnostic test currently available. However, there is a consilience of evidence pointing towards some sort of organic disease (possibly of a heterogeneous flavor) in at least a significant fraction of long covid patients. I doubt that all these biomarker findings are the fruit of indefatigable p-hacking but nevertheless, replicating the more promising findings would be nice (and essential to getting something into clinical use).
Research groups seem to favor hunting for *new* sets of exotic biomarkers (as seen by the ever-expanding lists of candidate biomarkers with good AUC in those "wtf *is* up with long covid" review articles) rather than trying to replicate existing biomarker findings or validate a more limited set of findings for clinical diagnostic use. I think this is unfortunate. There is a company claiming to do that with the Ang-1 and P-sel biomarkers summarized below, but they seem to be doing this alongside developing a novel platform ("quantum dot high performance fluorescent immunosensors for high sensitivity bio-detection and quantification to develop best-in-class low-cost rapid diagnostic assays") for point-of-care tests. Don't need high sensitivity detection and quantification to pick up on the vaporware vibes from this, and hopefully someone else tries to replicate the Ang-1 / P-sel findings, even if they can't commercialize it.
For instance, https://www.ncbi.nlm.nih.gov/pmc/articles/PMC9549814/ gets **excellent** classification accuracy -- between healthy subjects, people who've been in the ICU for covid, and long-covid outpatients -- by looking at blood levels of only two proteins; Angiopoietin-1 and P-selectin (both involved in endothelial / vascular / clotting function). If you look at the figures in that paper, you can see that the levels of Ang-1 and P-sel are **off the charts** in long covid outpatients compared to healthy patients -- or even people who've been in the ICU for covid and have recovered.
https://www.medrxiv.org/content/10.1101/2022.08.09.22278592v1 is another one of those "look at a bunch of immune-related biomarkers, throw statistics at it, and find what has good predictive accuracy" studies and yep, they indeed also find some biomarkers with excellent predictive accuracy. Also, the self-reported long-covid symptoms matched up well with the biomarker findings.
Looking a bit more upstream, this study found low levels of circulating Spike protein *months after acute illness* in 60% of studied long covid patients (and none in those who recovered OK from covid): https://academic.oup.com/cid/article/76/3/e487/6686531. The "weird viral persistence" stuff also shows up in this study of IBD patients -- almost all of which previously had a covid infection, but only half of which reported long covid symptoms: https://www.ncbi.nlm.nih.gov/pmc/articles/PMC9057012/. The majority of people with SARS-CoV-2 nucleocapsid protein and SARS-CoV-2 RNA in their guts reported long covid symptoms, and people without these covid virus parts in their gut reported no long covid symptoms.
I am very curious what people here make of long COVID biomarker findings. Multiple groups have found multiple sets of blood biomarkers, some with very good predictive value; see Table 2 in this review article https://www.thelancet.com/journals/ebiom/article/PIIS2352-3964(23)00117-2/fulltext.
Granted, the boundary between organic disease psychosomatic / functional neurological disorder might be demarcated by "is it detectable by well-validated clinical-grade diagnostics (in-vitro or otherwise) or not"; which I think is pure map/territory confusion. Under this definition, long covid is indeed psychosomatic; there is no such diagnostic test currently available. However, there is a consilience of evidence pointing towards some sort of organic disease (possibly of a heterogeneous flavor) in at least a significant fraction of long covid patients. I doubt that all these biomarker findings are the fruit of indefatigable p-hacking but nevertheless, replicating the more promising findings would be nice (and essential to getting something into clinical use).
Research groups seem to favor hunting for *new* sets of exotic biomarkers (as seen by the ever-expanding lists of candidate biomarkers with good AUC in those "wtf *is* up with long covid" review articles) rather than trying to replicate existing biomarker findings or validate a more limited set of findings for clinical diagnostic use. I think this is unfortunate. There is a company claiming to do that with the Ang-1 and P-sel biomarkers summarized below, but they seem to be doing this alongside developing a novel platform ("quantum dot high performance fluorescent immunosensors for high sensitivity bio-detection and quantification to develop best-in-class low-cost rapid diagnostic assays") for point-of-care tests. Don't need high sensitivity detection and quantification to pick up on the vaporware vibes from this, and hopefully someone else tries to replicate the Ang-1 / P-sel findings, even if they can't commercialize it.
For instance, https://www.ncbi.nlm.nih.gov/pmc/articles/PMC9549814/ gets **excellent** classification accuracy -- between healthy subjects, people who've been in the ICU for covid, and long-covid outpatients -- by looking at blood levels of only two proteins; Angiopoietin-1 and P-selectin (both involved in endothelial / vascular / clotting function). If you look at the figures in that paper, you can see that the levels of Ang-1 and P-sel are **off the charts** in long covid outpatients compared to healthy patients -- or even people who've been in the ICU for covid and have recovered.
https://www.medrxiv.org/content/10.1101/2022.08.09.22278592v1 is another one of those "look at a bunch of immune-related biomarkers, throw statistics at it, and find what has good predictive accuracy" studies and yep, they indeed also find some biomarkers with excellent predictive accuracy. Also, the self-reported long-covid symptoms matched up well with the biomarker findings.
Looking a bit more upstream, this study found low levels of circulating Spike protein *months after acute illness* in 60% of studied long covid patients (and none in those who recovered OK from covid): https://academic.oup.com/cid/article/76/3/e487/6686531. The "weird viral persistence" stuff also shows up in this study of IBD patients -- almost all of which previously had a covid infection, but only half of which reported long covid symptoms: https://www.ncbi.nlm.nih.gov/pmc/articles/PMC9057012/. The majority of people with SARS-CoV-2 nucleocapsid protein and SARS-CoV-2 RNA in their guts reported long covid symptoms, and people without these covid virus parts in their gut reported no long covid symptoms.