Objective: To assess the accuracy of traditional weaning indices in predicting extubation failure, and to compare their accuracy when indices are measured at the onset of a breathing trial (SBT) and at the end of the SBT before extubation.
Design: Prospective study.
Setting: Medical-surgical intensive care unit at a tertiary care hospital.
Patients: Four hundred eighteen consecutive infants and children who received mechanical ventilation for at least 48 h and were deemed ready to undergo a SBT by their primary physician.
Interventions: Respiratory frequency (RR), tidal volume (V(T)), maximal inspiratory pressure (P(imax)) and frequency-to-tidal volume ratio (f/V(T)) were obtained within the first 5 min of breathing through a T-piece. The primary physicians were unaware of those measurements and the decision to extubate a patient was made by them. RR, V(T), f/V(T) were remeasured before extubation by the respiratory therapists. Extubation failure was defined as needing re intubation within 48 h after extubation. The area under the receiver operating characteristic (ROC) curve was calculated for each index as a measure of the accuracy in predicting extubation outcome.
Measurements and main results: Three hundred twenty-three patients successfully underwent the SBT and were extubated, but 48 of them (14%) required re-intubation. The ROC curve for V(T), RR, P(imax) and f/V(T) measured within the first 5 min of breathing were 0.54, 0.56, 0.57 and 0.57, respectively. The ROC curve did not increase significantly when the above indices were remeasured before extubation.
Conclusions: In a population which had passed SBT, the ability of the traditional weaning indices to discriminate between children successfully extubated and children re-intubated is very poor.