Extubation Failure Prediction by Echography of the Diaphragm After Cardiothoracic Surgery: The EXPEDIA Study.

Abstract

Successful extubation is difficult to predict. Ultrasound measurement of the diaphragm thickening fraction (DTF) might help predict weaning failure after cardiothoracic surgery. We assessed the predictive performance of diaphragm ultrasound in a derivation cohort of 50 prospectively included cardiothoracic surgery subjects ready for a weaning trial and in a validation cohort of 39 subjects ventilated for ≥ 48 h. DTF was assessed by ultrasound during pressure support ventilation (PSV) then during a T-piece spontaneous breathing trial (SBT). DTF was the percentage change in diaphragm thickness between expiration and inspiration and DTFmax, the higher DTF value of the 2 hemidiaphragms. DTFmax during SBT (static study) and the difference in DTFmax between PSV and SBT (dynamic study) were analyzed. In the derivation cohort, DTFmax during SBT was 25.6 ± 17.3% in subjects with successful extubation and 65.2 ± 17.3% in those with weaning failure (difference 39.7 [95% CI 27.4-51.9], P < .01). During SBT, DTFmax ≥ 50% was associated with weaning failure (area under the receiver operating characteristic curve [AUC] 0.94 ± 0.05). In the dynamic study, a ≥ 40% DTFmax increase was associated with weaning failure (AUC 0.91 ± 0.06). In the validation cohort, DTFmax during SBT was 20.3 ± 17.1% in subjects with successful extubation and 82.0 ± 51.6% in those with weaning failure (difference 61.8 [95% CI 41.6-82.0], P < .01). During SBT, DTFmax ≥ 50% predicted weaning failure (AUC 0.99 ± 0.02). In the dynamic study, a ≥ 40% increase in DTFmax predicted weaning failure (AUC 0.81 ± 0.09). Measuring DTFmax during SBT and the DTFmax change when switching from PSV to SBT may help predict weaning failure after cardiothoracic surgery. The study was registered on ANZCTR. U1111-1180-1999.