Multimodal Ultrasound for Early Diagnosis and Prediction of ICU-Acquired Weakness

Abstract Background ICU-acquired weakness (ICU-AW) is a common long-term complication in critically ill patients whose diagnostic criteria are based on muscle strength scores and neuro-electrophysiological examination(e.g., electromyography,EMG). ICU-AW can lead to prolonged duration of mechanical ventilation, weaning difficulty, increased hospitalization costs, reduced long-term quality of life, and significant muscle atrophy, usually occurring on day 7 to 10 of hospitalization, so early recognition and intervention are critical for prognosis. However, current diagnostic is limited by variations in physician expertise, patient cooperation, the invasiveness and complexity of EMG, and environmental interference. Ultrasound, a simple, noninvasive, and safe imaging technique, has become critical in evaluating various diseases. This study aimed to investigate characteristics of multimodal ultrasound (two-dimensional and elastic ultrasound) in ICU-AW patients at different stages, and to identify potential diagnostic and predictive indicators. Methods This study enrolled 100 ICU inpatients from October 2023 to August 2024. The ultrasound parameters—cross-sectional area, muscle thickness, pinnate angle of the quadriceps, and subcutaneous fat thickness of both lower limbs—were recorded on Days 1, 5, and 10 of hospitalization. Muscle strength scores and/or EMG were performed on Day 10.The baseline, clinical and ultrasound data of the enrolled patients were analyzed and compared, the potential diagnoses and predictors were selected, and the model value was evaluated through the measures of diagnostic accuracy and the area under the ROC curve. Results A total of 100 patients were included in this study, including 43 patients underwent three complete ultrasound examinations,23 ICU-AW patients and 20 non-ICU-AW patients. (1) Univariate analysis showed that on the Day10, the cross-sectional area of rectus femoris muscle and the size of the feather angle of the rectus intermediate muscle in the ICU-AW group were significantly reduced compared with those in the non-ICU-AW group, which was statistically significant. (2) The change rate of left/right rectus femoris cross-sectional area (25.01%/29.66%, AUC = 0.87/0.89) and the change rate of pinnate angle of left/right vastus intermedius muscle (27.48%/27.56%, AUC = 0.81/0.84) were the best cutoff values. (3) The binary logistic regression model (based on age and pinnate angle of the femoral intermediate muscle on Day 5) had higher predictive performance (AUC = 0.889, sensitivity = 0.818, specificity = 0.813). Conclusion Using multimodal ultrasound measures, we developed a convenient, simple, and safe diagnostic index (the rate of change in ultrasound parameters between Day 1 and Day 10), confirming its potential as a standardized and efficient quantitative tool. In addition, we used binary logistic regression to establish a highly predictive model that may inform clinical decision-making and early intervention in ICU-AW.