Associations of serum glucose/potassium ratio with short-term and long-term mortality in sepsis patients: a retrospective cohort study based on the MIMIC-IV database

This study evaluated the prognostic value of glucose/potassium ratio (GPR) in sepsis using MIMIC-IV data (n = 31,717). Patients were stratified by GPR quartiles (Q1: <1.450 to Q4: ≥2.169). Adjusted Cox models demonstrated a dose-dependent increase in mortality in patients with higher GPR levels (P < 0.001), with the highest risk of mortality observed in Q4. Restricted cubic spline analysis revealed a U-shaped relationship between GPR and mortality (30-day optimal cutoff: 1.49). Sensitivity analyses excluding patients with diabetes, acute renal failure, or chronic liver disease confirmed the robustness of results, with consistent findings across subgroups. These findings indicate that GPR independently predicts both short-term (30/90/180-day) and long-term (1-year) mortality in sepsis, providing valuable references for risk stratification.