ABSTRACT
Background: Neurocognitive symptoms are common in acute as well as convalescent (post-acute sequelae of COVID-19 [PASC]) COVID-19, but mechanisms of CNS pathogenesis are unclear. The aim of this study was to investigate cerebrospinal fluid (CSF) biomarker evidence of CNS infection, immune activation and neuronal injury in convalescent compared with acute infection. Method(s): We included 68 (35% female) patients >=18 years with CSF sampled during acute (46), 3-6 months after (22) SARS-CoV-2 infection or both (17), and 20 (70% female) healthy controls from longitudinal studies. The 22 patients sampled only at 3-6 months were recruited in a PASC protocol. CSF N-Ag was analyzed using an ultrasensitive antigen capture immunoassay platform (S-PLEX SARS-CoV-2 N Kit, Meso Scale Diagnostics, LLC. Rockville, MD). Additional analyses included CSF beta2-microglobulin (beta2M)], IFN-gamma, IL-6, TNF-alpha neurofilament light (NfL), and total and phosphorylated tau. Log-transformed CSF biomarkers were compared using ANOVA (Tukey post-hoc test). Result(s): Patients sampled during acute infection had moderate (27) or severe (19) COVID-19. In patients sampled at 3-6 months, corresponding initial severity was 10 (mild), 14 (moderate), and 15 (severe). At 3-6 months, 31/39 patients reported neurocognitive symptoms;8/17 patients also sampled during acute infection reported full recovery after 3-6 months. CSF biomarker results are shown in Figure 1. SARS-CoV-2 RNA was universally undetectable. N-Ag was detectable only during acute infection (32/35) but was undetectable in all follow up and control samples. Significantly higher CSF concentrations of beta2M (p< 0.0001), IFN-gamma (p=0.02), IL-6 (p< 0.0001) and NfL (p=0.04) were seen in acute compared to post-infection. Compared to controls, beta2M (p< .0001), IL-6 (p< 0.0001) and NfL (p=0.005) were significantly higher in acute infection. No biomarker differences were seen post-infection compared with controls. No differences were seen in CSF GFAp, t-tau or p-tau. Conclusion(s): We found no evidence of residual infection (RNA, N-Ag), inflammation (beta2M, IL-6, IFN-gamma, TNF-alpha), astrocyte activity (GFAp) or neuronal injury (NfL, tau) 3-6 months after initial COVID-19, while significantly higher concentrations of several markers were found during acute infection, suggesting that PASC may be a consequence of earlier injury rather than active CNS damage. CSF beta2M, IL-6, IFN-gamma and NfL were significantly lower after 3-6 months than during acute COVID-19 and not different from healthy controls. (Figure Presented).
ABSTRACT
Background: The underlying CNS pathogenesis in COVID-19 is not clear and viral RNA is rarely detected in cerebrospinal fluid (CSF). We measured viral antigen and biomarker profiles in CSF in relation to neurological symptoms and disease severity. Methods: We included 44 (32% female) hospitalized patients (26 moderate, 18 severe COVID-19) and 10 healthy controls (HC). 21 patients were neuroasymptomatic (NA), 23 neurosymptomatic (NS;encephalopathy=21, encephalitis=1, GBS=1). For antigen and cytokine analyses, a patient control (PC;n=41) group (COVID-negative with no sign of CNS infection in clinical CSF samples) was used. CSF nucleocapsid antigen (N-Ag) was analyzed using an ultrasensitive antigen capture immunoassay platform, S-PLEX direct detection assay, S-PLEX SARS-CoV-2 N Kit (MesoScale Diagnostics, LLC. Rockville, MD). Additional analyses included CSF neopterin, β2-microglobulin, cytokines and neurofilament light (NfL). Results: CSF N-Ag was detected in 31/35 patients (0/41 controls) while viral RNA was negative in all. CSF N-Ag was significantly correlated with CSF neopterin (r=0.38;p=0.03) and IFN-γ (r=0.42;p=0.01) adjusted for sampling day. No differences in CSF N-Ag concentrations were found between patient groups. All patient groups had markedly increased CSF neopterin, β2M, IL-6, IL-10 and TNF-α compared to controls, while IL-2, IL-1β and IFN-γ were significantly increased only in the NS group. CSF biomarkers were associated with time from symptom onset to CSF sampling. After adjusting for time of sampling, the NS group had significantly higher CSF IFN-γ (p=0.03), and showed a statistical trend towards significantly higher CSF neopterin, IL-6 and TNF-α (p=0.056-0.06) than the NA group. Additionally, age-adjusted CSF NfL was higher in the NS compared to the HC (p=0.01) group. No differences were seen in any CSF biomarkers in moderate compared to severe disease. Conclusion: Viral antigen is detectable in CSF in a majority of patients with COVID-19 despite the absence of detectable viral RNA, and is correlated to CNS immune activation markers. Patients with neurological symptoms had a more marked immune activation profile compared to NA patients, as well as signs of neuroaxonal injury compared to controls. These observations could not be attributed to a difference in COVID-19 severity. Our results highlight the importance of neurological symptoms and indicate that the CNS immune response and CNS pathogenesis can be initiated by viral components without direct viral invasion of the CNS.