Risk Factors and Prediction Model for ICU-Acquired Weakness in Severe Acute Pancreatitis: A Retrospective Cohort...

To investigate the incidence and independent risk factors of intensive care unit-acquired weakness (ICU-AW) in patients with severe acute pancreatitis (SAP) and to develop and validate a corresponding risk prediction model.

In this retrospective study, 277 SAP patients admitted to the ICU between January 2020 and March 2024 were enrolled. ICU-AW was diagnosed using the Medical Research Council sum score (MRC-SS), with a score <48 defining the condition. Patients were randomly split into a modeling cohort (70%) for model development and a validation cohort (30%) for internal validation. Univariate and multivariate logistic regression analyses were employed to identify factors associated with ICU-AW. A nomogram was constructed based on independent predictors identified from the multivariate analysis. The model's performance was assessed by the area under the receiver operating characteristic curve (AUC) and the Hosmer-Lemeshow (H-L) test.

The incidence of ICU-AW was 38.6% (107/277) among the enrolled SAP patients. Multivariate analysis identified age ≥51 years (OR=5.83), APACHE II score ≥14 (OR=11.19), ICU length of stay ≥7 days (OR=12.61), use of neuromuscular blocking agents (OR=3.34), D-lactate ≥41.69 mg/L (OR=2.92), and presence of nutritional risk (OR=5.34) as independent risk factors for ICU-AW. Early rehabilitation care was a significant protective factor (OR=0.18). The prediction nomogram incorporating these seven factors demonstrated excellent discrimination, with an AUC of 0.922 (95% CI: 0.884-0.959) in the modeling cohort and 0.918 (95% CI: 0.873-0.963) in the validation cohort. The H-L test indicated good calibration in both cohorts (P=0.903 and P=1.000, respectively).

ICU-AW is a common complication in SAP patients. We developed a nomogram that effectively integrates readily available clinical variables (age, disease severity, ICU stay, medication use, D-lactate, nutritional status, and rehabilitation care) to individually predict the risk of ICU-AW. This model may assist clinicians in early identification of high-risk patients and facilitate targeted preventive strategies.