Ultra-processed foods and diet quality in long COVID: associations with symptom burden, DNA damage, and inflammation.

Diet quality can be a potential modifier of post-acute COVID-19 syndrome (PACS), yet the relationship between dietary patterns, symptom burden, and biological markers remains poorly understood. We hypothesized that a low-quality diet would positively associate with PACS symptom burden, DNA damage, inflammatory and neurotrophic markers, and with overall health quality. A total of 186 individuals were classified into three groups: control (n = 64), acute PACS (n = 66), and chronic PACS (n = 55). Dietary intake was assessed by a food frequency questionnaire, and diet quality was quantified via the Protective and Risk Eating Score (PRES). Blood samples were analyzed for inflammatory (tumor necrosis factor-α, transforming growth factor-β, interleukin-6, interleukin-1β) and neurotrophic (brain-derived neurotrophic factor, glial cell line-derived neurotrophic factor, S100 calcium-binding protein B) markers. DNA damage was assessed by micronuclei frequency, telomere length, and comet assay parameters. Health index (HI) was assessed by daily habits (alcohol consumption, sleeping patterns, hydration, exercise frequency, smoking habits, anthropometric parameters, and PRES). No significant differences were found among groups in ultra-processed foods or sugary products intake, PRES scores, inflammatory or neurotrophic biomarkers, DNA damage indicators, and HI. Of note, poorer diet quality was associated with higher symptom burden in PACS women, mainly in fatigue, memory, mental health, circulatory, and skin symptom domains. Few associations among diet quality, HI, and biomarkers emerged. Of note, principal component analysis shows that PACS relies on the convergence of genomic instability, systemic inflammation, and low diet quality. These findings suggest that diet may represent a modifiable factor in PACS management, warranting longitudinal and interventional studies to determine causality and therapeutic potential.