Integrated ultrasound protocol in predicting weaning success and extubation failure: a prospective observational study

Abstract

Background: Difficulty in weaning from mechanical ventilation is encountered in approximately 20% of patients in the intensive care unit. We assessed the utility of a combined lung, diaphragmatic, and cardiac ultrasound protocol to predict extubation failure. Methods: All patients extubated following a successful spontaneous breathing trial (SBT) were included in the study. Lung ultrasonography score (LUS), diaphragmatic thickness fraction (DTF), changes in velocity time integral (VTI) to passive leg raise at the beginning of SBT, and change in LUS following SBT were recorded. Results: A total of 60 patients who underwent successful SBT were included in the study. Twenty-seven patients required either non-invasive or invasive mechanical ventilation during the next 48 hours and were classified as weaning failure (Group F). The remaining 33 patients were designated as weaning success (Group S). Compared to group S, patients in Group F had significantly longer ICU length of stay (6.96 ± 4.30 days vs. 11.66 ± 3.85 days, P < 0.001), higher LUS change during SBT (1 [0–2] vs. 2 [1–4], P < 0.001), lower DTF (30.87 ± 5.32 vs. 27.88 ± 6.24, P = 0.04), and showed lower VTI increment to PLR (13.63 ± 3.44 vs. 9.11 ± 4.59, P < 0.001). Using a binary logistic regression model, DTF < 26% (odds ratio 6.20, 95% CI: 1.06–36.04) and VTI change to PLR < 10.2% (odds ratio 6.16, 95% CI: 1.14–33.13) were found to be significant predictors of weaning failure (P < 0.05). The AUROC for VTI and DTF for predicting weaning failure were 0.79 and 0.64, respectively. Conclusions: An integrated ultrasound protocol using a combination of lung, diaphragm, and cardiac sonography was a reliable predictor of weaning failure.