Immune dysregulation and endothelial dysfunction associate with as determinants of a pro-thrombotic profile thrombotic risk in Long COVID

Alicia Simón Rueda¹Clara Sánchez-Menéndez¹Guiomar Casado¹Daniel Fuertes²MARIA-ARANZAZU MURCIANO-ANTON³Elena Mateos¹Susana Domínguez- Mateos⁴Francisco Pozo¹Javier García-Pérez¹Mayte Pérez¹Miguel Cervero⁵Marta Massanella⁶Gemma Moncunill⁷Montserrat Torres¹Mayte Coiras^1*

• ¹Carlos III Health Institute (ISCIII), Madrid, Spain
• ²Escuela Técnica Superior de Ingeniería de Telecomunicación, Universidad Politécnica de Madrid., Madrid, Madrid, Spain
• ³Centro de Salud Pedro Laín Entralgo, Madrid, Asturias, Spain
• ⁴Centro de Salud Arroyomolinos, Arroyomolinos, Spain
• ⁵Fundación Universidad Alfonso X el Sabio (FUAX), Madrid, Madrid, Spain
• ⁶IrsiCaixa, Barcelona, Catalonia, Spain
• ⁷Instituto Salud Global Barcelona (ISGlobal), Barcelona, Catalonia, Spain

Long COVID (LC) affects approximately 10% of individuals post-SARS-CoV-2 infection, with symptoms persisting beyond 12 weeks. The underlying mechanisms remain unclear, and current models often focus on pre-existing comorbidities. This cohort study aimed to identify robust biomarkers and clarify LC pathogenesis through a comprehensive analysis performed in 32 LC individuals 26 months post-infection compared with 35 fully recovered individuals recruited between March and July 2022. Blood and fecal samples were collected, and multiple parameters associated with immune dysfunction, endothelial damage, bacterial translocation, and coagulation alterations, alongside signs of viral persistence and sociodemographic and clinical features, were analyzed. Although viral RNA was undetected on blood or stool, elevated plasma IgG against the nucleocapsid may indicate frequent reinfections, greater infection severity, or delayed immune normalization. Increased levels of prothrombin, thrombin, fibrinogen, sEPCR, and CRP pointed to persistent endothelial dysfunction and coagulation imbalance. Lower levels of the bactericidal protein REG3A suggest potential disruptions in gut microbiota regulationmucosal immune response. We found no major differences in traditional comorbidities, highlighting that LC may stem from distinct pathogenic mechanisms beyond pre-existing conditions. Importantly, our study revealed impaired humoral immunity and identified an association between vaccine heterogeneity and increased LC risk, emphasizing the relevance of consistent vaccination strategies. A Random Forest model using the measured biomarkers achieved 100% accuracy in classifying LC individuals, reinforcing their diagnostic and prognostic potential. These findings support a multifactorial model of LC involving immune dysregulation and persistent endothelial damage that led to coagulation abnormalities and a pro-thrombotic profileincreased thrombotic risk, supporting that LC is more closely related to a sustained, uncontrolled inflammatory response rather than immunodeficiency, and underscoring the value of multidimensional biomarker profiling for guiding clinical management and prevention strategies.