Association between the FT3/FT4 ratio and in-hospital mortality among hospitalized patients undergoing thyroid function testing

Background Thyroid function is highly sensitive to acute illness and systemic stress in hospitalized patients, and the free triiodothyronine/free thyroxine (FT3/FT4) ratio has been proposed as an integrative marker reflecting peripheral thyroid hormone metabolism. This study aimed to investigate the associations between thyroid function parameters, particularly the FT3/FT4 ratio, and in-hospital all-cause mortality. Methods A total of 26,223 inpatients from all departments of a large tertiary hospital in China were included. Serum thyroid-stimulating hormone (TSH), FT3, FT4, and the FT3/FT4 ratio (×100) were measured at admission, and in-hospital all-cause mortality was identified from electronic medical records. Cox proportional hazards regression models were used to assess associations between thyroid hormone parameters and mortality. Nonlinear associations were evaluated using restricted cubic spline analyses. Sensitivity analyses included exclusion of extreme thyroid hormone values, inverse-probability-of-selection weighting using the entire source population, and stratified Cox regression to address non-proportional hazards. Subgroup and exploratory analyses of admission vital signs were also performed. Results Higher TSH (HR = 1.015, 95% CI: 1.010–1.020) and FT4 (HR = 1.017, 95% CI: 1.002, 1.033) were associated with higher in-hospital mortality, whereas higher FT3 (HR = 0.865, 95% CI: 0.806, 0.929) and FT3/FT4 ratio (×100) (HR = 0.974, 95% CI: 0.964, 0.983) were associated with lower mortality. RCS analyses showed evidence of nonlinearity for TSH, but not for FT3, FT4, or the FT3/FT4 ratio. The inverse association for the FT3/FT4 ratio remained in the IPW analysis (HR = 0.954, 95% CI: 0.945, 0.963) and in the Cox model stratified to address non-proportional hazards (HR = 0.970, 95% CI: 0.961, 0.980). Sensitivity analyses excluding extreme values generally supported the primary findings, particularly for FT3 and the FT3/FT4 ratio. Subgroup analyses suggested potential heterogeneity across several patient characteristics. Exploratory analyses showed associations between the FT3/FT4 ratio and admission vital signs, whereas only temperature was independently associated with mortality after multivariable adjustment. Conclusion The FT3/FT4 ratio was independently associated with in-hospital mortality and may serve as a potential indicator of outcomes among patients undergoing thyroid function testing.