Beyond Euthyroid Sick Syndrome: Pre-Admission Levothyroxine Use and Thyroid Hormone–Outcome Associations in Hospitalized Patients

This is the authors' abstract. We don't add key points for this paper.

Background/Objectives: Pre-admission levothyroxine use may influence the interpretation of thyroid tests beyond the conventional framework of non-thyroidal illness syndrome (NTIS). We examined whether thyroid hormone–outcome associations differed by treatment status. Methods: This retrospective study included medical inpatients with thyroid-stimulating hormone (TSH) < 5 mIU/L during 2006–2024. Propensity-score matching used age, sex, and 16 comorbidity indicators. Complete-case negative-binomial and logistic models examined hospitalization length and in-hospital mortality, with treatment-by-hormone interactions. Results: Among 14,332 matched patients, 3717 had complete data (1986 users; 1731 nonusers). Higher free triiodothyronine (FT3) in nonusers was associated with shorter hospitalization (incidence–rate ratio [IRR] 0.76, 95% confidence interval [CI] 0.73–0.80) and lower mortality odds (odds ratio [OR] 0.47, 95% CI 0.35–0.62). Free thyroxine (FT4) associations differed by levothyroxine use for both hospitalization length (interaction IRR 0.95, 95% CI 0.94–0.97) and mortality (interaction OR 0.83, 95% CI 0.77–0.90; both p < 0.001). At mean hormone concentrations, levothyroxine use was associated with shorter hospitalization (IRR 0.87, 95% CI 0.81–0.94) and lower mortality odds (OR 0.53, 95% CI 0.35–0.82). Crude mortality was 4.6% versus 5.0%, respectively; the conditional estimates represent a different comparison. Conclusions: FT4–outcome associations differed by pre-admission levothyroxine use, extending the established low-T3 prognostic framework to treatment-specific hormone patterns. These findings support considering treatment history when interpreting inpatient thyroid tests. Selective testing and residual confounding limit causal interpretation.