Development of a nomogram to predict hospitalized mortality of patients with sepsis

Background: Sepsis is a severe organ dysfunction caused by infection, which is a leading cause of morbidity and mortality in critically ill patients. We developed a novel risk-predicted nomogram of hospitalized mortality for this retrospective study. Methods The study involved a retrospective analysis of sepsis patients diagnosed between January 2013 and December 2019. Based on clinical outcomes, patients were categorized into a survival group and a death group. Using logistic regression analysis, a predictive model with a nomogram was formulated in the training set, after which internal validation and sensitivity analysis were performed. C-index and calibration curves were used in the training and validation cohorts to evaluate nomogram discrimination and calibration. Utilizing decision curve analysis, an assessment of the clinical utility of the final nomogram was conducted. Results A total of 494 patients were enrolled for analysis. In the multivariate analysis, cardiac function (OR: 1.58, P = 0.039) stood out as an independent sepsis risk factor, whereas hemoglobin (OR: 0.98, P = 0.025) served as a protective factor. Although the multivariate regression analysis did not detect any significant differences in AKI staging, in light of the previous studies, a nomogram for the prediction of hospitalized sepsis mortality was constructed by taking into account these three factors. Based on this model's C-index of 0.626, it suggested moderate predictive power. Conclusions Our study indicates that patients with sepsis in combination with anemia, cardiac dysfunction, and AKI are associated with significantly high mortality. To some extent, this may contribute to the determination of risk assessments and treatment strategies.