Abstract
Altered thyroid function (TF), especially non-thyroidal illness syndrome (NTIS), is common in hospitalised COVID-19 patients, but adaptive responses to disease as captured by blood biomarkers in the context of thyroid signal integration require further elucidation. This retrospective study analysed 29 routine biomarkers, including immune, inflammatory, haematological, organ damage, and TF markers, from patients admitted to the University Hospital Coventry and Warwickshire, decoding associations of the biomarkers with TF key determinants and identifying patterns of severity-specific thyroid dysfunction. Patients were categorised by disease severity based on admission to either critical care units (CCU) or general wards. Among 237 records, 140 were CCU admissions (mean age 57, 69% male) and 97 were general ward patients (mean age 63, 59% male). Both groups exhibited negative correlations between haematology and TF biomarkers with inflammation and organ dysfunction markers. The pattern of clustering and the deduced network built around the correlation of TF components also exhibited disease severity-specific characteristics. At least 11 biomarkers, including free-triiodothyronine (fT3) and free-thyroxine (fT4), showed significant group-specific differences. Notably, CCU patients demonstrated an interleukin-6-dependent enhancement of correlations between TSH and fT3 or the fT3/fT4 ratio. Our findings suggest an immune-driven adaptation of the pituitary-thyroid axis, often presenting as NTIS in COVID-19. Deciphering information from routine biomarkers could enhance TF assessment, particularly in critically ill patients.