The prediction model of septic shock was constructed with multiple parameters
10.3760/cma.j.issn.1671-0282.2025.06.017
- VernacularTitle:多参数构建脓毒性休克预测模型
- Author:
Liling LAI
1
;
Baoquan CHEN
;
Zhu'e WU
;
Jiguang ZHOU
;
Huobiao SHE
;
Ming CHEN
Author Information
1. 福建医科大学附属漳州市医院超声医学科，漳州 363000
- Keywords:
Multi-parameter;
Septic shock;
Left ventricular pressure-strain loop;
Myocardial work;
Global longitudinal strain;
Brain natriuretic peptide;
Lactic acid;
Me
- From:
Chinese Journal of Emergency Medicine
2025;34(6):852-857
- CountryChina
- Language:Chinese
-
Abstract:
Objective:To develop and validate a predictive model for septic shock using noninvasive left ventricular pressure-strain loop myocardial work parameters combined with clinical indicators.Methods:In this retrospective study, 105 septic patients admitted to the intensive care unit of Zhangzhou Hospital between January and December 2023 were analyzed after screening (initial cohort: 124 patients). Participants were stratified into non-shock ( n=38) and shock ( n=67) groups based on septic shock occurrence within 48 hours of admission. Demographic characteristics, laboratory parameters, clinical variables, and myocardial work indices were compared between groups. Independent risk factors were identified through multivariate logistic regression, and a predictive model was constructed. Results:Five independent predictors of septic shock were identified: Left ventricular global longitudinal strain (GLS) ≥ -16% (X 1).Global work index (GWI) <1 196.5 mmHg% (X 2).Brain natriuretic peptide (BNP) ≥299.81 pg/mL (X 3).Lactate (Lac) ≥4.75 mmol/L (X 4).Mean arterial pressure (MAP) <68.5 mmHg (X 5).The derived model equation was:Y = -21.104 + 3.517×X 1 + 2.066×X 2 + 1.941×X 3 + 3.440×X 4 + 3.526×X 5.ROC analysis determined an optimal diagnostic cutoff of -13.6185 points (rounded to -14 for clinical practicality). Scores≥-14 indicated high septic shock risk. The model demonstrated excellent discrimination (AUC=0.960, 95% CI=0.919-0.998) and goodness-of-fit (Hosmer-Lemeshow test, P=0.804). Conclusions:This novel predictive model integrating myocardial work parameters and clinical indicators exhibits outstanding diagnostic performance for early septic shock detection, potentially enabling timely therapeutic intervention.
