Unraveling the Complex Web of Long COVID symptoms: A Network Analysis Approach

Abstract

Long COVID is a syndrome of persistent symptoms following a COVID-19 infection. With over 200 identified symptoms and ongoing debate about the central symptoms of this condition, this study aimed to explore the network structure of self-reported long COVID symptoms. Participants with persistent symptoms after a COVID-19 infection (n=458 , Mage = 36.0±11.9; 46.5% male) indicated the prevalence and severity of their long COVID symptoms using a comprehensive 114-item symptom list spanning across 10 health domains. Correlation and regularized partial correlation network analyses were performed to examine the interrelatedness of symptoms focusing on symptoms listed in international health guidelines and those with a prevalence of at least 20% in our sample. Replication analyses were performed using a second dataset from a representative Dutch sample (n=415). We find that some long COVID symptoms are closely interconnected (e.g., congestion and neurocognitive symptoms), whereas others (e.g., fatigue and dizziness) are more dispersed between clusters. Centrality indices indicated that the most common symptoms mentioned in the guidelines were not necessarily the most central in the networks. These findings were largely replicated in the second dataset. The study provides a visualization of the wide array of long COVID symptoms, revealing it as a diverse syndrome marked by different clusters of loosely connected symptoms. Although explorative, the results can be informative for potential pathogenic pathways involved in long COVID-19 symptoms and may inform future research and clinical interventions.