There’s a new study out from the NIH (open access) tracking a cohort of people who became infected with SARS-CoV-2, and following them for an extended period. This sort of longitudinal data collection was set up to look for post-infection symptoms (generally known as “long Covid” or “post-Covid”), or in the paper’s own acronym, PASC (for Postacute Sequelae of SARS-CoV-2 Infection). We really do need more hard data on this topic, because there have been a great many reports of people who have taken much longer than they expected to feel normal after a bout of the disease. The correlation between PASC and severity/duration of the initial infection, with factors like age and gender and underlying physical conditions, with various forms of treatment (or with various iterations of the coronavirus itself!) are so far basically unknown. We really do need to understand what’s going on, what the eventual public health burden might be, and what we can do about it. As someone who’s just recently recovered from the damn virus myself, I find the topic of even greater than normal interest.

This paper is going to be controversial. Actually, it already is controversial, and news of its findings has just been circulating in the last day or so. You’ll see why momentarily, but here’s the setup: the clinicians (starting in June of 2020) started enrolling people who had been diagnosed with coronavirus (some of them hospitalized during their illness), and a group of antibody-negative control patients. The enrollment is continuing, but this paper reports on the earliest cohorts, 189 post-corona patients (12% of whom were hospitalized before recovering), and 120 control patients. Patients in the former group who had symptoms and confirmed lab diagnosis of Covid-19 had to be at least six weeks past the onset of their infection, with no fever in the previous seven days. Asymptomatic patients had to be at least four weeks past their RT-PCR diagnosis. The mean time from onset to enrollment was 162 days. Control group participants were recruited as they showed up, with no filters for age, medical conditions, etc., and they were not matched to the other group in these variables, although the two groups were overall quite similar.

PASC was defined as any symptom that developed (or worsened) with the infection and was still present at enrollment, weeks later. Only symptoms that showed up in at least 1% of the study participants were counted. For control patients, only symptoms that began or worsened during the previous 162 days were counted, to keep the evaluation period as similar as possible compared to the post-Covid patients. Everyone had a baseline evaluation (medical history, blood work, etc.) Blood work was taken at intervals during the study, as well as measurement of cardiovascular function, pulmonary function, and cognitive function (using the NIH Toolbox set of assays). The neurocognitive results were normalized for age, etc., as is usual with that assay suite. Participants also took standard question batteries about quality of life, anxiety, depression, physical condition, etc.

OK, the results. The people in the post-Covid group definitely reported more symptoms than the control group. The most common of these were fatigue, difficulty breathing, altered sense of smell or taste, difficult concentrating, and headaches. Also high on the list were memory problems, chest pain, insomnia, anxiety, general body pains, and tinnitus. Physical examinations were much less conclusive: the only difference that was clear between the two groups was tenderness or pain in bursae, muscles, or tendons in the post-corona patients. The blood work was even less differentiated: levels of C-reactive protein (general inflammation marker), D-dimer (clotting problems), troponin-1 (heart problems), pro-B type natriuretic peptide (also a heart marker), and neurofilament light chain (marker of nerve degeneration) were all statistically indistinguishable between patients and controls. Similarly, standard markers of kidney and liver function were indistinguishable. The study also looked for signs of autoimmune disease, such as anti-nuclear antibodies, anticardiolipin antibodies, and rheumatoid factor (anti-IgG antibodies), but these also showed no difference. There were a few positive tests for these in each group, but in every case, none of the patients involved showed any signs of lupus, rheumatoid arthritis, or other autoimmune syndromes. A subset of people from each group got an even more thorough immunological workup (macrophage inflammatory protein-1β, interferon-γ, TNF-α, PD-L1, interferon γ–induced protein 10 (CXCL10), interleukin-2 receptor α, interleukin-1β, interleukin-6, interleukin-8, RANTES (CCL5), and CD40), but again, no differences were detected. Examination of T-cell populations (CD4+, CD8+, and more) showed a difference only in CD4+ cluster 6 cells, which were higher in post-Covid patients, but (again) turned out not to be associated with reports of PASC. Only two of the post-Covid patients showed the presence of the viral N protein, one with PASC and one without. So there was no evidence of persistent viral infection, or of immune system dysfunction.

There was no statistical difference in pulmonary function (spirometry, lung volume, etc.), and no statistical differences in any of the cardiovascular tests (echocardiograms, etc.) But the median distance walked in a six-minute walking test was lower for the post-Covid patients as compared to the controls (560 meters versus 595). But even then, the changes in walking distance did not correlate with PASC symptoms. The only variables in patient histories and so on that correlated with a greater likelihood of showing PASC were being female or having a self-reported history of anxiety disorders, but as you can see from the above, nothing really tracked at the biochemical level. No differences were found in the neurocognitive tests (processing speed, episodic memory, executive function) between the two groups, nor was any correlation apparent with reports of PASC. Self-reported quality of life was indeed different, though, with lower scores in people who also reported PASC symptoms. A standard anxiety evaluation (GAD-2) showed a signifigant correlation between higher anxiety scores and reports of symptoms as well.

See, I told you this was going to be controversial! As mentioned, this is an ongoing study, but so far there are no diagnostic findings that would allow you to even say for sure that post-Covid even exists, biochemically. The authors certainly don’t put it that way, but I will. And I have to say that I’m quite surprised, as well as perhaps a bit relieved. But at the same time, there are a lot of people who most certainly have reported post-Covid symptoms, so there’s more work to be done here. We shouldn’t be ready just to wave these patients away, but at the same time, we’re going to have to deal with the physiological implications of this study, which argue against most of the plausible medical causes.

The authors mention the limitations of their study, chief of which is that the majority of the patients studied post-Covid did not have severe disease. A similar study among people who went through a much harder coronavirus experience would be more likely to pick up differences, you would certainly think. But balancing against that are the reports that post-Covid symptoms have not seemed to correlate (anecdotally) with the severity of infection. They also mention that they may have actually tested more PASC-reporting patients than you might expect in the general population, because these people may have been more likely to enroll in the study. PASC that resolved earlier would not have been caught, of course, and finally, this is only 189 patients. I look forward to further data from this study, and from others, but this one does put a lot of markers down. If there is a physiological/biochemical/immunological cause for “long Covid”, we have not seen it yet. This was a very thorough search, but it turned up nothing in those categories. . .