Longitudinal trajectories of neurologic symptoms and cognitive associations at 36-months post-hospitalization for COVID-19.

BACKGROUND: A substantial proportion of individuals hospitalized for COVID 19 experience persistent neurological and psychiatric symptoms. The long-term trajectories of these symptoms, collectively referred to as neuropsychiatric manifestation of Post-Acute Sequelae of COVID-19 (Neuro-PASC), and their association with cognitive outcomes remains poorly understood. We conducted a 36-month longitudinal cohort study to characterize Neuro-PASC symptom trajectories, examine their associations with domain specific cognitive performance, and identify baseline biological predictors of long-term neurological outcomes. METHODS: We performed a longitudinal observational study of 227 adults ages 19-79 (mean = 47.6), previously hospitalized for COVID-19 symptoms. Participants completed follow-up assessments at 3/6-, 12-, 24-, and 36-months post-hospitalization. At each visit, structured, self-reported neurological symptom surveys, digital cognitive testing (BrainCheck), and updated medical history were collected. Generalized estimating equation models assessed longitudinal associations between cognitive subdomains and individual symptoms. K-means clustering identified distinct symptom trajectory groups. Baseline demographic and clinical factors were examined and compared across symptom trajectory groups. RESULTS: Several neurological symptoms increased in prevalence over 36 months following hospitalization in some individuals. Fatigue remained most common (53-65%), followed by memory/concentration difficulties (47-61%) and sleep disturbances (40-48%). Dizziness had a modest increase (30-49%), while coordination impairment more than doubled (18-41%). Headaches remained prevalent and relatively stable (21-35%). Symptom co-occurrence strengthened over time, with fatigue frequently clustering with other symptoms (up to 45% at 36 months). Overall BrainCheck composite scores were not independently associated with neurological symptoms, but significant associations emerged for executive function, memory, and attention, including time-dependent effects. Trajectory analyses identified distinct longitudinal patterns (1. stable trajectory, 2. worsening trajectory, and 3. improving trajectory), associated with baseline demographics, comorbidities, and acute disease severity. CONCLUSION: Neurological symptoms can persist or worsen up to 36 months post- COVID-19 hospitalization, even as some cognitive functions show partial recover. Distinct symptom trajectories and evolving associations with specific cognitive domains and biological outcomes underscore the heterogeneous and dynamic nature of Neuro-PASC, supporting the need for long-term neuropsychological monitoring and informing targeted interventions and mechanistic studies.