What question did this study set out to answer?

This research aims to develop and validate a hemodynamic scoring system for fluid resuscitation in acute pancreatitis patients.

February 21, 2026BMC Gastroenterology0 citationsOpen Access

Development of a hemodynamic scoring system for acute pancreatitis: an exploratory approach to optimizing fluid resuscitation

GHGuanhao HuangGuangxi Medical University HHHuali HuangGuangxi Medical University

Key Points

This research aims to develop and validate a hemodynamic scoring system for fluid resuscitation in acute pancreatitis patients.
Extracted clinical data from eICU 2.0 and MIMIC-IV 3.1 databases.
Patients were stratified into death and survival groups based on their outcomes.
Key predictors were identified using LASSO regression and multivariable logistic regression.
A nomogram prediction model and cyclic scoring system were developed.
Model performance was evaluated using ROC curves and other statistical methods.
The scoring system achieved an AUC of 0.80 and 0.81 for the eICU and MIMIC-IV databases, respectively.
Average arterial pressure, age shock index, urea nitrogen, and potassium levels were the key predictors.
The model demonstrated good discriminatory power and clinical utility in guiding fluid resuscitation.
It effectively identified high-risk acute pancreatitis patients for clinical intervention.

Abstract

Our study aims to establish and validate a scoring framework based on hemodynamic parameters in patients with acute pancreatitis (AP), and to evaluate its clinical predictive performance and significance in guiding fluid resuscitation. The clinical data of AP patients admitted to the ICU for the first time were extracted from the eICU 2.0 and MIMIC-IV 3.1 databases. Subjects were stratified into death and survival groups based on in-hospital outcomes. Key predictors were identified via intergroup comparison, LASSO regression, and multivariable logistic regression, and were used to develop a nomogram prediction model and a cyclic scoring system. Model performance was evaluated via receiver operating characteristic (ROC) curves, calibration plots, Hosmer–Lemeshow test, decision curve analysis, and internal and external validation. The hemodynamic scoring model was constructed by imitating the Framingham risk scoring system, and its clinical utility for fluid resuscitation was assessed via ROC analysis, box plots, and line graphs. This study included 707 patients (42 deaths and 665 survivors) from the eICU 2.0 database and 423 patients (38 deaths and 385 survivors) from the MIMIC-IV 3.1 database, all admitted to the ICU for AP. After screening, average arterial pressure (MAP), age shock index (ASI), urea nitrogen (BUN), and potassium ion (K+) were selected to construct a prediction model. The constructed scoring system was internally and externally validated using the eICU 2.0 database and MIMIC-IV 3.1 database, with an area under the curve (AUC) of 0.80 (95% CI: 0.74–0.88) and 0.81 (95% CI: 0.72–0.89), respectively. When applied to fluid resuscitation in AP patients, this model demonstrated potential utility in guiding therapy. The clinical model established based on MAP, ASI, BUN and K+ demonstrates good discriminatory power, calibration, and clinical utility in predicting the mortality rate of AP patients in the ICU. The established hemodynamic scoring system can identify high-risk AP patients and guide clinical intervention, and has potential value in guiding fluid resuscitation for AP patients in the ICU.

KI fragen

Bookmark

View Full Paper