947 – Post-Acute SARS-CoV-2 Antigenemia Is Associated With Some but Not All Long COVID Symptoms, 2025, Peluso et al

947 – Post-Acute SARS-CoV-2 Antigenemia Is Associated With Some but Not All Long COVID Symptoms

Michael Peluso, Thomas Dalhuisen, Zoe Swank, Sarah Goldberg, Scott Lu, Samuel Han, Jessica Chen, Emily Fehrman, Khamal Anglin, Rebecca Hoh, Timothy J. Henrich, Dan Kelly, Steven G. Deeks, David Walt, Jeffrey Martin

Background
SARS-CoV-2 (SCV2) RNA and antigens can be found in several tissues many months after infection, but whether SCV2 persistence causes Long COVID (LC) is unproven. The scant data in favor of causation has related persistent SCV2 to non-specific LC definitions. It is unknown whether persistent SCV2 is associated with all, or only some, of the different symptoms of this heterogenous condition.

Methods
Sampling from UCSF’s LIINC cohort, we defined 5 LC case groups and 1 non-LC control group. Consistent with the NASEM definition, LC groups had ≥1 symptom in a given grouping that was rated “very bothersome” at ≥2 visits 3-14 months after the first confirmed SCV2 infection. LC case groups were: a) neurologic (headache, brain fog or dizziness); b) cardiopulmonary (cough, dyspnea, chest pain or palpitations); c) gastrointestinal (GI; nausea, vomiting, diarrhea, constipation, belly pain or appetite loss); d) musculoskeletal (back, joint or muscle pain); and e) fatigue. The non-LC group was those whose acute symptoms resolved within 30 days of infection and who had no symptoms 3-14 months post-infection. SCV2 spike, S1, and nucleocapsid (N) antigens were measured in plasma by single molecule array (Simoa) assay. Odds ratios (OR) were used to relate antigenemia at any visit 3-14 months post-infection to concurrent symptoms.

Results
Spike antigenemia was found in 7.5% (3/40) of the non-LC group. In unadjusted analyses, spike antigenemia was associated with neurologic, cardiopulmonary, and musculoskeletal symptoms but not GI or fatigue (Table). After adjustment for confounding, associations endured for neurologic (OR 4.5, p=0.037), cardiopulmonary (OR 4.8, p=0.08), and musculoskeletal (OR 5.5, p=0.049) symptoms. After exclusion of timepoints within 1 month of known SCV2 vaccination or reinfection, point estimates were similar but confidence intervals widened slightly. No significant associations were seen with S1 or N antigens.

Conclusions
Among adults with prior SCV2 infection, we found evidence for an association between SCV2 spike antigenemia 3-14 months post-infection and some but not all LC symptoms. This work highlights the need for well-characterized participants, with and without a variety of symptoms, to yield associations, and it suggests that different mechanisms might contribute to the variable presentation of LC. Nonetheless, given the novelty of the system, more work is needed (including intervention studies) to confirm the findings and exclude confounding before causation is inferred.

Link (Conference on Retroviruses and Opportunistic Infections) [Abstract Only]