Serial C-Reactive Protein Measurements for Prognostication and Antimicrobial Decision-Making in ICU Sepsis Patients: A Retrospective Cohort Study.

Background: C-reactive protein (CRP) is a widely used biomarker for sepsis management; however, its prognostic value remains uncertain. This study evaluated whether serial CRP measurements are associated with mortality in intensive care unit (ICU) patients with sepsis. Methods: This retrospective cohort study included adult patients admitted to a tertiary-care ICU between 2020 and 2023, who had at least three serial CRP measurements around days 1, 3, and 7 to guide sepsis management as diagnosed by the treating team. We examined the relationship between CRP levels and 90-day mortality. Results: The study included 171 patients (median age: 67 years; 55% male). Compared to patients with lower CRP levels, those with day-1 CRP ≥75 mg/L (n=87, 50.9%) had similar clinical characteristics, except for a higher prevalence of malignancy. Antibiotic use and clinical outcomes were similar between the groups. Meropenem duration tended to be shorter in patients whose CRP levels decreased from days 1 to 7 than in those whose CRP levels increased (median 7 days vs. 10 days; p=0.06). 42 patients (24.6%) died within 90 days. Compared to survivors, non-survivors had slightly higher CRP levels on day 1 (91 mg/L vs. 70 mg/L, p=0.08) and similar changes in CRP over time. Receiver operating characteristic curve analysis demonstrated that both CRP levels and their temporal changes had poor prediction for 90-day mortality (area under the curve for day-1 CRP, 0.590), with day-1 procalcitonin performing slightly better (area under the curve, 0.658). In the multivariable logistic regression analysis, day-1 CRP was not associated with 90-day mortality. Conclusion: Serial CRP measurements did not reliably predict mortality in ICU patients with sepsis but may have influenced the antibiotic duration. Routine CRP monitoring in ICU patients with sepsis may have a limited prognostic value.