Persistent fatigue in long-COVID is not associated with peripheral inflammatory or cellular stress biomarkers: A cross-sectional controlled study

Background: Fatigue persists as a dominant and debilitating phenomenon in long-COVID, yet its underlying biological mechanisms remain unclear. While inflammatory variables tend to normalize within months post-infection, fatigue continues to significantly impact quality of life. Understanding whether specific biomarkers associate with long-COVID fatigue could shed light on pathophysiological mechanisms and potential therapeutic targets. Methods: In this single-center, cross-sectional controlled study, we enrolled 48 individuals with long-COVID (according to NICE criteria) and 48 age- and sex-matched recovered controls with prior SARS-CoV-2 infection but no persistent symptoms. We carefully excluded all subjects with other diseases or conditions that could influence fatigue levels. Fatigue severity was assessed using three validated instruments: Fatigue Visual Analog Scale (fVAS), Functional Assessment of Chronic Illness Therapy-Fatigue (FACIT-F), and SF-36 vitality subscale. Blood samples were analyzed for pro-inflammatory markers (CRP, TNF-α, IL-6, IL-1β) and biomarkers associated with cellular stress responses and neuroprotection (HSP90α, APOA4, Serpin F1/PEDF, Hemopexin). Anti-nuclear antibodies (ANA) were tested to assess potential autoimmune mechanisms. Depression was assessed using the Hospital Anxiety and Depression Scale, Depression Subscale (HADS-D). Results: Long-COVID patients demonstrated significantly higher fatigue severity across all instruments compared to recovered controls: fVAS median scores 63 versus 5 (p < 0.001), FACIT-F scores 21.5 versus 49 (p < 0.001), and SF-36 vitality scores 25 versus 72.5 (p < 0.001). Depression scores were also significantly elevated in long-COVID cases. However, none of the measured biomarkers differed significantly between groups: HSP90α, Serpin F1, Hemopexin, APOA4, and CRP showed no differences, while TNF-α and IL-6 showed only tendencies toward higher levels in long-COVID (p = 0.07 and p = 0.07, respectively). IL-1β concentrations were in most cases below the lower limit of detection and were excluded from further analysis. ANA positivity was 10.4% in cases versus 4.2% in controls (p = 0.38) and did not influence fatigue levels. Multivariable regression analysis revealed no significant associations between biomarkers and fatigue severity. Conclusions: Fatigue in long-COVID represents severe, persistent disability comparable to observations in chronic inflammatory diseases and chronic fatigue syndrome but is not associated with traditional inflammatory biomarkers or cellular stress response proteins measured in peripheral blood. The absence of biomarker associations suggests that long-COVID fatigue may involve more complex mechanisms, potentially including persistent neuro-immune dysregulation, epigenetic changes, or pathophysiological processes not reflected in systemic biomarker concentrations including neurobiological mechanisms such as altered predictive processing and central nervous system-confined neuroinflammation. These findings highlight the need for alternative approaches to understanding and treating long-COVID fatigue beyond conventional inflammatory paradigms.