Combined lung and diaphragm ultrasound predicts extubation outcomes in ARDS: a prospective study.

Background: Extubation failure is a crucial issue for acute respiratory distress syndrome (ARDS). Ultrasound of the lung and diaphragm is individually valuable for predicting extubation outcomes. We aimed to determine whether combined lung and diaphragmatic ultrasound could improve the accuracy of predicting the extubation of ARDS patients.
Methods: This was a prospective cohort study of ARDS patients who were ready for extubation. The lung ultrasound score (LUS), diaphragmatic displacement (DD), diaphragm thickening fraction (DTF), and diaphragmatic-rapid shallow breathing index (D-RSBI) were measured at the end of the spontaneous breathing trial. The primary outcome was extubation success. Logistic regression was used to combine these indicators, and the predictive performance of the single and combined indicators was evaluated through receiver operating characteristic (ROC) curves, the Hosmer-Lemeshow Ĉ-test, and the Brier score. Multivariate logistic regression was used to determine the association between combined ultrasound indicators and extubation success.
Results: This study enrolled 132 eligible patients from January 2019 to December 2022. A total of 71% (94/132) of patients were successfully extubated from mechanical ventilation. The combination of LUS and D-RSBI had the largest area under the ROC curves, the lowest Brier score, and the greatest calibration. After formula transformation, LUS + 2.43 × D-RSBI ≤ 14.273 was significantly associated with extubation success in ARDS patients.
Conclusions: In ARDS patients receiving mechanical ventilation, the combination of LUS and D-RSBI was more accurate than a single parameter alone in predicting extubation outcomes. This combined approach could help refine extubation protocols in critical care. Clinical trial registration This study is registered online with the Chinese Clinical Trial Registry (ChiCTR), http://www.chictr.org.cn , ChiCTR1800019340 (Registration time: 2018/11/06).
(© 2024. The Author(s).)