SESSION TITLE: Novel Assessment and Treatments for Respiratory Failure
SESSION TYPE: Original Investigation Slide
PRESENTED ON: Tuesday, October 27, 2015 at 11:00 AM - 12:15 PM
PURPOSE: To test the hypothesis that Ultrasonographically (US) measured Diaphragm thickness (DT) is a good predictor of successful extubation.
METHODS: Design and setting: A prospective observational study in medical intensive care unit between 1st July to 28th February 2015. Methodology: Patients planned for extubation were included in the study. Pregnancy, age<18yrs and those with poor ultrasound windows were excluded. High frequency linear probe was used to measure the right sided DT at the zone of apposition(ZOA) between 8th to 10th intercostal spaces in mid-axillary line, just before extubation. Logistic regression was used to develop a model with extubation failure as the outcome and change in DT(delta fraction), Rapid shallow Breathing index (RSBI), and other patient covariates as predictors. Assuming a re-intubation rate of 25% and four explanatory variables, a sample size of two hundred was thought appropriate. Extubation failure was defined as re-intubation within 48 hrs of extubation.
RESULTS: In the logistic model, delta fraction was the only significant predictor of extubation success among all covariates (const 8.895,Coef.-0.2915, p value 0.00). The model showed very good discrimination (receiver operating curve, ROC area of 0.95) but poor calibration (Hosmer-Lemeshow chi2(3) = 208.53, Prob > chi2 = 0.000).
CONCLUSIONS: Ultrasonographic measurement of DT at the ZOA just before extubation is a good predictor of extubation success.
CLINICAL IMPLICATIONS: Ultrasonographic measurement of the diaphragm thickening at ZOA just before extubation is a good predictor of extubation success or failure.Being easily available and repeatable,diaphragm ultrasound can be used as a bedside tool to assess readiness for extubation.Further research regarding ultrasonographic diaphragmatic evaluation are needed to assess this technique in a greater number of patients with various diseases.
DISCLOSURE: The following authors have nothing to disclose: Alai Taggu
No Product/Research Disclosure Information