To investigate the predictive value of lung ultrasound combined with diaphragmatic ultrasound for extubation failure in patients with acute respiratory distress syndrome (ARDS).

A total of 126 ARDS patients admitted to our hospital from January 2022 to October 2025 were enrolled. Among them, 99 patients with successful extubation were assigned to the control group, and 27 patients with extubation failure were assigned to the observation group. Bedside ultrasonography was performed 4 hours after the completion of a spontaneous breathing trial. The lung ultrasound (LUS) score was calculated using the 12zone method, and diaphragmatic excursion and diaphragmatic thickening fraction were measured. Clinical data and ultrasound parameters were compared between the two groups. LASSO regression and multivariate logistic regression were used to analyze risk factors for extubation failure, and receiver operating characteristic (ROC) curves were plotted to evaluate the predictive value of each indicator.

Compared with the control group, the observation group had a higher median LUS score (15.00 vs. 9.00, P<0.001), and lower values of quiet breathing diaphragmatic excursion (QDE) (P=0.027), deep breathing diaphragmatic excursion (DDE) (P=0.006), quiet breathing diaphragmatic thickening fraction (QDTF) (P<0.001), and deep breathing diaphragmatic thickening fraction (DDTF) (P<0.001). Multivariate logistic regression analysis showed that sequential organ failure assessment (SOFA) score (OR=1.713), oxygenation index (OR=0.991), duration of mechanical ventilation (OR=1.078), LUS score (OR=1.686), and DDTF (OR=0.872) were independent risk factors for extubation failure in ARDS patients (all P<0.05). The area under the curve (AUC) of the combination of LUS score and DDTF for predicting extubation outcome was 0.923 (95%CI: 0.861~0.963), which was superior to that of LUS score alone [0.817 (95%CI: 0.732~0.876), P=0.003] or DDTF alone [0.812 (95%CI: 0.732~0.876), P=0.009]. ROC curve analysis of the combined model including SOFA score, oxygenation index, duration of mechanical ventilation, LUS score, and DDTF showed that the combined clinical and bedside ultrasound model achieved an AUC of 0.969 (95%CI: 0.921~0.991), with a sensitivity of 88.89% and a specificity of 93.94%. The HosmerLemeshow goodnessoffit test indicated good calibration (χ²=2.448, P=0.964). Clinical decision curve analysis showed that when the highrisk threshold ranged from 0.03 to 0.95, the combined prediction of extubation failure provided a high net clinical benefit.

Bedside multi-organ ultrasound monitoring can serve as an important assessment tool for extubation in patients with ARDS. By combining bedside lung ultrasound with diaphragmatic ultrasound, together with the patient′s clinical symptoms and signs, this approach can effectively guide the selection of extubation timing, help improve the extubation success rate, and is of clinical significance in ensuring that ARDS patients are extubated smoothly and safely.