Construction and validation of a prediction model for short-term mortality risk in elderly patients with sepsis: a retrospective study.

BACKGROUND

This study aims to investigate the risk factors affecting short-term mortality in elderly patients with sepsis and a nomogram model was constructed.

METHODS

A retrospective collection was performed on clinical data of 1116 elderly patients (≥ 65 years) with sepsis hospitalized from January 2014 to December 2023. Patients were divided based on 30-day mortality: death group (392 cases) and survival group (724 cases). Differences in various biochemical indicators were compared, and Cox regression analysis was used to identify independent risk factors. The model's performance was evaluated using receiver operating characteristics (ROC) and Kaplan-Meier curves.

RESULTS

Univariate and multivariate Cox survival analysis were utilized to explore the risk factors for short-term mortality in elderly patients with sepsis. The analysis exhibited a significant association of age, procalcitonin, antiarrhythmic drugs, vasopressors, and albumin with patient survival time. Group differences in vitamin C positivity were observed, but this variable was not included in the prediction model. The nomogram model, demonstrated high diagnostic accuracy (C-index 0.790). Patients were divided into the high-risk group and the low-risk group based on the model's thresholds, showing significantly shorter survival times for the high-risk group.

CONCLUSION

Increased age, use of antiarrhythmic drugs and vasopressors are independent risk factors for mortality in elderly patients with sepsis, whereas high albumin levels are protective factors. The developed nomogram provides an effective tool for predicting short-term mortality in this patient population.