Dynamic changes in regional ventilation distribution and end-expiratory lung impedance using electrical impedance tomography: associations with weaning and extubation outcomes

Abstract

Background Accurately assessing the optimal timing for extubation remains a significant challenge in intensive care units (ICUs). The rapid shallow breathing index (RSBI) is widely used but lacks insight into regional lung physiology. This study aimed to characterize dynamic changes in regional ventilation distribution and end-expiratory lung impedance (EELI) using electrical impedance tomography (EIT) during spontaneous breathing trials (SBTs) and to explore their association with weaning and extubation outcomes. Methods In this prospective observational study, 52 mechanically ventilated patients undergoing their first SBT were enrolled. EIT parameters, including the Global Inhomogeneity (GI) index, center of ventilation (COV), ventrodorsal ventilation asymmetry (|ROI V-D|), and changes in EELI (ΔEELI), were recorded at four time points. Weaning failure was defined as SBT failure, or as reintubation or noninvasive ventilation within 48 h after extubation. Extubation failure was defined as reintubation or rescue NIV within 48 h. Logistic models coded 1 = failure and 0 = success. Results SBT failure (n = 19, 36.5%) was associated with progressive deterioration of ventilation, with a significant increase in the GI index (p < 0.001) and a notable spike in |ROI V-D| at T3 (p = 0.005). Extubation failure (n = 9, 27.3% of SBT successes) was associated with a greater cumulative loss of end-expiratory lung volume (ΔEELI T4–T1) compared with the success group (p < 0.001). In secondary exploratory analyses, a combined MLE model of T4 RSBI and T3 GI showed an apparent AUC of 0.923 (95% CI 0.840–0.986) for SBT failure (optimism-corrected AUC: 0.905). A combined model of T4 RSBI and ΔEELI (T4–T1) showed an apparent AUC of 0.875 (95% CI: 0.688–0.988) for extubation failure; with only 9 events, this estimate is hypothesis-generating and requires external validation. Conclusions EIT-derived parameters offer insights into regional ventilation changes during the weaning process. Increased ventilation inhomogeneity and ventrodorsal mismatch are associated with SBT failure, while loss of end-expiratory lung impedance is associated with extubation failure. Adding EIT parameters to RSBI provided exploratory incremental information for discriminating SBT failure and extubation failure in this cohort. These secondary analyses require external validation before any clinical application can be considered. Clinical Trial Registration: https://www.chictr.org.cn, identifier ChiCTR2500108219.