Objective: The development of weaning predictors in mechanically ventilated children has not been sufficiently investigated. The purpose of this study was to evaluate the accuracy of some weaning indices in predicting weaning failure.
Design: Prospective, interventional study.
Setting: University-affiliated children's hospital with a 19-bed intensive care unit.
Patients: 84 consecutive infants and children requiring mechanical ventilation for at least 48 h and judged ready to wean by their primary physicians.
Interventions: Patients who met the criteria to start weaning underwent a trial of spontaneous breathing lasting up to 2 h. Bedside measurements of respiratory function were obtained immediately before discontinuation of mechanical ventilation and within the first 5 min of spontaneous breathing. The primary physicians were blinded to those measurements, and the decision to extubate a patient at the end of the spontaneous breathing trial or reinstitute mechanical ventilation was made by them. Failure to wean was defined as the requirement for mechanical ventilation at any time during the trial of spontaneous breathing (trial failure) or needing reintubation within 48 h of extubation (extubation failure).
Measurements and main results: Seventy-five patients had neither signs of respiratory distress nor deterioration in gas exchange during the trial and were extubated. Twelve patients required reintubation within 48 h. In 9 patients, mechanical ventilation was reinstituted after a median duration of the spontaneous breathing trial of 35 min. The only independent predictor of trial failure was tidal volume indexed to body weight [odds ratio 2.60, 95 % confidence interval (CI) 1.40 to 24.9]. The only independent predictor of extubation failure was frequency-to-tidal volume ratio indexed to body weight (odds ratio 1.23, 95 % CI 1.11 to 1.36). The sensitivity, specificity, and positive and negative predictive values to predict weaning failure were calculated for each of the above variables. These values were 0.48, 0.86, 0.53, and 0.83, respectively, for a frequency-to-tidal volume ratio higher than 11 breaths/min per ml per kg and 0.43, 0.94, 0.69, and 0.83, respectively, for a tidal volume lower than 4 ml/kg.
Conclusions: Three-quarters of ventilated children can be successfully weaned after a trial of spontaneous breathing lasting 2 h. Both tidal volume and frequency-to-tidal volume ratio indexed to body weight were poor predictors of weaning failure in the study population.