Can Low Cortisol Predict Long COVID? A Controversial Issue

Cortisol dysregulation has been proposed as a biomarker of long COVID (LC), but findings remain inconsistent. Prior reports suggested low cortisol levels in LC, yet collection times and study designs varied substantially. To evaluate morning serum cortisol distributions in an independent LC cohort, accounting for circadian timing and sleep patterns, we performed a retrospective cross-sectional study of consecutive adults seen at the Stanford Long COVID Clinic between 14 February 2022 and 31 July 2024 (IRB #62996). Eligible participants had confirmed SARS-CoV-2 infection, symptoms persisting ≥3 months per NASEM criteria, completion of the Alliance Sleep Questionnaire (ASQ), and a morning serum cortisol measured using the Roche Elecsys^® Cortisol II assay. Analyses were restricted to collections between 05:00-10:00, categorized as early morning peak (EMP: 05:00-08:00) or mid-morning (MMP: 08:01-10:00). Cortisol was classified as low (<6.2 μg/dL), normal (6.2-19.4 μg/dL), or elevated (>19.4 μg/dL). Among 86 LC patients (69.8% female; mean age 45.4 ± 12.9 years), the mean serum cortisol level was 15.67 ± 6.76 μg/dL. Overall, 62.8% of patients had cortisol within the reference range, 36.0% had elevated levels, and only 1.2% (n = 1) had a low value. Cortisol distributions were comparable across the EMP and MMP collection windows, with no statistically significant differences observed by sleep alignment. Inflammatory markers, including CRP and D-dimer, were largely within reference ranges across all cortisol strata. Contrary to earlier reports, low morning cortisol was rare in this LC cohort; most values were normal or elevated. Findings underscore the importance of circadian timing when interpreting cortisol in LC and highlight the need for prospective studies with serial measurements to determine biomarker utility.