Can electrical impedance tomography predict successful mechanical ventilation weaning in tracheostomized patients? A single-center observational study

Abstract

Objective: To evaluate whether ventilation variables monitored by electrical impedance tomography (EIT) behaved differently between tracheostomized patients who progressed successfully or failed within 6 h after withdrawal of invasive mechanical ventilation (IMV), as well as to improve the accuracy of failure prediction models. Methods: This was a single-center prospective observational study including tracheostomized adults who had been on IMV for more than seven days and who were clinically ready for discontinuation. Clinical and EIT variables—end-expiratory lung impedance, impedance variation, center of ventilation, heterogeneity index, and ventilation distribution—were collected. The primary outcome measure was IMV weaning failure within 6 h, defined as the need to resume IMV for respiratory distress. Associations between EIT variables and failure were tested by using multivariate logistic regression and ROC curves to assess predictive accuracy. Results: Fifty-nine patients were analyzed; 43 (72.9%) were successfully weaned from IMV, and 16 (27.1%) failed. In the region of interest H1 (ventral lung region), end-expiratory lung impedance decreased more markedly among successfully weaned patients (p = 0.03). Respiratory rate behavior differed between the groups over time (interaction p = 0.01), whereas other EIT variables showed no group-time interaction. In the multivariate analysis, BMI and 24-hour fluid balance were independently associated with failure. Adding EIT variables to clinical models did not improve predictive accuracy. Conclusions: The EIT variables evaluated in our study did not behave differently between tracheostomized patients who were successfully weaned from IMV and those who were not. The addition of EIT variables did not improve the accuracy of failure prediction models.