Seen some debate on this, so thought I’d share some points to consider:
After reading the new gut-biopsy paper and old Peluso paper, my view is that the evidence supports long-term persistence of SARS-CoV-2 material in gut tissue, but it still does not prove that this is specific to long COVID or that it is the main driver of symptoms. In the new preprint, spike RNA and protein were detected in long COVID cases, but also in controls. The main between-group signal was that colon spike RNA was higher in long COVID, with higher RNAscope H-score and higher average spike signal per cell in colon. But importantly, spike protein was not higher in that colon comparison, and there was also no significant difference in colon or ileal orf1ab-sense signal. So the summary is that they found more colon spike RNA, but not more spike protein-positive cells, in long COVID versus controls.
The transcriptomic comparison is also weaker than the headline suggests. In the colon Spike ROI comparison, after multiple-testing correction the only surviving genes were Y-chromosome or male-specific genes. The paper itself states that the long COVID cohort was predominantly female while the control cohort was predominantly male. So one of the main surviving gene-level signals may simply reflect sex imbalance rather than long COVID biology. The paper also did not clearly show that the local spike-associated abnormalities tracked with symptom burden, including GI symptom status.
There is also an important limitation in the paper’s “local environment” argument. The strongest mechanistic analysis, where they look at Spike versus Spike- regions, was then done mainly within the long COVID group itself. That is where they report macrophage, monocyte, plasma cell, and Treg enrichment, along with correlations with antigen-presentation, inflammation, and complement-related genes such as C1QA, C1QB, and C1QC. The abstract says healthy-control Spike versus Spike- colon tissue showed a more modest response with 38 DEGs, but the paper does not present the same depth of follow-up for controls in the main results. So we still do not know whether spike-positive regions in controls might also sit in a similar local immune niche.
Peluso’s paper has a similar limitation. It found spike single-stranded RNA in rectosigmoid tissue in all five biopsied participants and double-stranded RNA in three, up to 676 days after infection. But all five biopsied participants had long COVID symptoms, so that biopsy arm cannot tell us whether this is more common in long COVID than in recovered people without persistent symptoms.
So my thoughts are more: these studies support that persistence is real and may matter in some patients, but they still do not prove that it is the main explanation for long COVID overall. My interpretation remains that persistence may drive disease in a subset, while in others the illness may become partly self-sustaining through other changes.
Outside of long COVID, the evidence increasingly suggests that SARS-CoV-2 can persist and be linked to immune changes in a broad range of people, including so-called healthy controls. What that means for the general population over time remains an open question.
But that is not the same as showing that viral persistence is the unique driver of long COVID, or that clearing it would cure long COVID. That remains unproven. It may be true for a subset of patients, but the current studies do not establish it. Too often, key steps are missing from the evidence, yet the findings get discussed as though the issue is settled. It is not.
So very likely yes spike = bad, but whether spike = long Covid, different question that as it stands has no solid answer I think. If anything we know for sure that spike on its own does not cause long covid. Does spike in a very specific spot cause it, does the quantity cause it, does it only cause it for a subgroup? All still completely unanswered I feel.
A new preprint examines gut biopsies from people with LongCOVID and healthy controls. It does not just ask whether SARS2 Spike is present in tissue, but also what is happening in the surrounding tissue using spatial transcriptomics. That is probably the most interesting part of the paper.🧵