Impact of Methylprednisolone on Short- and Long-Term Outcomes in Critically Ill Patients with Sepsis: A Cohort Study Using the MIMIC-IV Database

Background Low-dose hydrocortisone reduces mortality in sepsis, but the dose–response profile of methylprednisolone (MP) remains unclear. We assessed the independent associations of the initial dose, cumulative dose, and frequency of methylprednisolone administration with 30-day and 1-year mortality in critically ill adults with sepsis. Methods This retrospective cohort study included 9,876 adults meeting Sepsis-3 criteria from the MIMIC-IV database (2008–2022). Exposures were categorized as: initial dose (< 100 mg vs ≥ 100 mg), cumulative dose (< 100 mg vs ≥ 100 mg), and number of MP administrations (1, 2–4, 5–9, ≥ 10). Cox proportional-hazards models, Boruta feature selection, and extensive sensitivity analyses were used to estimate adjusted hazard ratios (HRs) with 95% confidence intervals (CIs). Results MP exposure increased 30-day mortality by 23% (HR 1.23, 95% CI 1.10–1.38) and 365-day mortality by 26% (HR 1.26, 95% CI 1.15–1.38). Risk plateaued when the initial or cumulative dose reached ≥ 100 mg. A single administration carried the highest risk (30-day HR 1.48, 365-day HR 1.39). Two-to-four administrations also sustained elevated risk (30-day HR 1.20, P = 0.045; 365-day HR 1.25, P = 0.001). Five-to-nine administrations yielded point estimates < 1, but 95% CIs included unity (30-day HR 0.84, P = 0.27; 365-day HR 0.99, P = 0.94), indicating no significant benefit. ≥10 administrations significantly reduced only 30-day mortality (HR 0.30, 95% CI 0.10–0.93) without improving 365-day survival (HR 1.08, 95% CI 0.70–1.66). Boruta ranked MP as the important predictor of death. All sensitivity analyses were consistent. Conclusion Intravenous MP is independently associated with increased 30-day and 1-year mortality in sepsis; risk plateaus at initial or cumulative doses ≥ 100 mg. Frequency analyses show single-bolus administration is most harmful, 5–9 low-dose administrations confer no significant benefit, and ≥ 10 administrations only short-term reduce 30-day mortality without improving long-term survival. When MP is clinically necessary, we recommend the “three-low” rule: low initial dose, low daily dose, low total cumulative dose, with reassessment for prompt discontinuation within 5–9 days. Multicenter prospective cohort studies are warranted to refine daily dosing and indications to minimize sepsis mortality.