Organism type of infection is associated with prognosis in sepsis: an analysis from the MIMIC-IV database

BACKGROUND: Sepsis has a high mortality rate, which is expensive to treat, and is a major drain on healthcare resources; it seriously impacts the quality of human life. The clinical features of positive or non-positive blood cultures have been reported, but the clinical features of sepsis with different microbial infections and how they contribute to clinical outcomes have not been adequately described. METHODS: We extracted clinical data of septic patients with a single pathogen from the online Medical Information Mart for Intensive Care(MIMIC)-IV database. Based on microbial cultures, patients were classified into Gram-negative, Gram-positive, and fungal groups. Then, we analyzed the clinical characteristics of sepsis patients with Gram-negative, Gram-positive, and fungal infections. The primary outcome was 28-day mortality. The secondary outcomes were in-hospital mortality, the length of hospital stay, the length of ICU stay, and the ventilation duration. In addition, Kaplan–Meier analysis was used for the 28-day cumulative survival rate of patients with sepsis. Finally, we performed further univariate and multivariate regression analyses for 28-day mortality and created a nomogram for predicting 28-day mortality. RESULTS: The analysis showed that bloodstream infections showed a statistically significant difference in survival between Gram-positive and fungal organisms; drug resistance only reached statistical significance for Gram-positive bacteria. Through univariate and multivariate analysis, it was found that both the Gram-negative bacteria and fungi were independent risk factors for the short-term prognosis of sepsis patients. The multivariate regression model showed good discrimination, with a C-index of 0.788. We developed and validated a nomogram for the individualized prediction of 28-day mortality in patients with sepsis. Application of the nomogram still gave good calibration. CONCLUSIONS: Organism type of infection is associated with mortality of sepsis, and early identification of the microbiological type of a patient with sepsis will provide an understanding of the patient's condition and guide treatment.