The diaphragm-intercostal index for predicting weaning failure in mechanically ventilated patients: a prospective cohort study.

Background: Accurately predicting the risk of weaning failure is crucial for optimizing clinical decision-making and improving patient prognosis. Despite widespread use, the predictive value of diaphragm ultrasound remains controversial, and the compensatory role of auxiliary respiratory muscles has been overlooked. This study aims to evaluate whether the diaphragm-intercostal index (DII), a novel ultrasound index for joint assessment of diaphragm and intercostal muscle function, can predict weaning failure more accurately than traditional parameters. Methods: This prospective single-center cohort study enrolled invasive mechanical ventilation patients (≥48 h) ready for weaning. Ultrasound was performed 30 min into the spontaneous breathing trial (SBT) (pre-extubation, for primary predictive analysis) and 24 h post-extubation (for exploratory descriptive analysis only). Measurements included diaphragmatic excursion (DE), diaphragm thickening fraction (TFdi), and intercostal muscle thickening fraction (TFic). DII was calculated as TFic/TFdi. The primary outcome was weaning failure, defined as SBT failure or need for therapeutic non-invasive ventilation (NIV), reintubation, or death within 48 h post-extubation. Sensitivity analysis was performed after excluding SBT failure patients to assess DII's performance in predicting true post-extubation failure. Lasso-logistic regression and ROC curves assessed predictive performance. Results: Of 107 patients, 81 weaned successfully and 26 failed. Among failures, 5 (19.23%) had SBT failure, 8 (30.77%) required reintubation, 11 (42.31%) required therapeutic NIV, and 2 (7.69%) died. DII's AUC was 0.952 (95% CI: 0.907-0.998), significantly outperforming RSBI (AUC = 0.789) and TFdi (AUC = 0.830). The optimal DII cutoff was >0.485 (sensitivity 80.8%, specificity 97.5%). In the sensitivity analysis excluding SBT failures, DII's AUC for predicting true post-extubation failure remained high at 0.942 (95% CI: 0.887-0.996), and DeLong test showed no significant difference from the primary analysis ( Conclusion: DII can effectively quantify the compensation of the auxiliary respiratory muscles relative to the diaphragm, showing favorable predictive performance for predicting weaning failure, strikingly superior to traditional parameters. However, validation in large-scale, multi-center studies is needed, particularly after standardizing SBT protocols with positive end-expiratory pressure (PEEP) as per guidelines, and in populations including obese and heart failure patients.