Antibodies against SARS-CoV-2 spike protein in the cerebrospinal fluid of COVID-19 patients and vaccinated controls: a multicentre study

Introduction
SARS-CoV-2 antibodies in the cerebrospinal fluid (CSF) of COVID-19 patients possibly reflect blood-cerebrospinal fluid barrier (BCB) disruption due to systemic inflammation. However, some studies indicate that CSF antibodies signal a neurotropic infection. Currently, larger studies are needed to clarify this, and it is unknown if CSF antibodies appear solely after infection or also after COVID-19 vaccination. Therefore, we aimed to investigate the CSF dynamics of SARS-CoV-2 antibodies in a multicenter study of COVID-19 patients and vaccinated controls.

Methods
A cohort study of Danish and Norwegian COVID-19 patients and controls investigated with a lumbar puncture (April 2020–December 2022). Serum and CSF were analysed locally for routine investigations, and centrally at Statens Serum Institut (Danish governmental public health institute) for SARS-CoV-2 IgG antibodies against the spike protein using the Euroimmun (quantitative) and Wantai (qualitative) assays. Primary outcome was the quantity of CSF SARS-CoV-2 antibodies post-COVID versus post-vaccination. Secondary outcomes included regression models examining the relationship between CSF antibodies and serum levels, albumin ratio, CSF pleocytosis, COVID-19 severity, and temporal antibody dynamics.

Results
We included 124 individuals (Mean [SD] age 47.2 [16.6]; 59.7% males surviving COVID-19 and controls. Of these, 86 had paired CSF-serum testing. Antibody-index calculations did not support a SARS-CoV-2 brain infection. Multi-variate regression revealed that CSF SARS-CoV-2 antibodies were most strongly influenced by serum antibody levels and BCB permeability, as measured by increasing albumin ratio. CSF antibody levels displayed a dose–response relationship (p < 0.0001) influenced by preceding vaccinations or infections. CSF antibody levels (median [IQR]) were highest among those both previously infected and vaccinated, 100.0 [25.0–174.0], and those vaccinated without prior infection, 85.0 [12.0–142.0], and lowest among previously infected individuals without preceding vaccination, 5.9 [2.7–55.1], (p = 0.003). SARS-CoV-2 antibodies in CSF were also detected via qualitative assays in the COVID-19 (46.8%) and vaccinated (78.6%) groups, p = 0.03.

Conclusion
SARS-CoV-2 antibodies detected in CSF can be derived following both infection and vaccination for COVID-19. CSF antibody levels increase in a dose–response relationship with the number of prior infections and vaccinations and are most strongly influenced by serum antibody levels and BCB permeability. These findings stress the importance of carefully interpreting CSF antibody results when assessing neurological complications following infections not categorized as neurotropic.