Objective: To determine whether the predictive accuracy of clinical judgment alone can be improved by supplementing it with an objective weaning protocol as a decision support tool. Methods: This was a multicenter prospective cohort study carried out at three medical/surgical ICUs. The study involved all consecutive difficult-to-wean ICU patients (failure in the first spontaneous breathing trial [SBT]), on mechanical ventilation (MV) for more than 48 h, admitted between January of 2002 and December of 2005. The patients in the protocol group (PG) were extubated after a T-piece weaning trial and were compared with patients who were otherwise extubated (non-protocol group, NPG). The primary outcome measure was reintubation within 48 h after extubation. Results: We included 731 patients-533 (72.9%) and 198 (27.1%) in the PG and NPG, respectively. The overall reintubation rate was 17.9%. The extubation success rates in the PG and NPG were 86.7% and 69.6%, respectively (p < 0.001). There were no significant differences between the groups in terms of age, gender, severity score, or pre-inclusion time on MV. However, COPD was more common in the NPG than in the PG (44.4% vs. 17.6%; p < 0.001), whereas sepsis and being a post-operative patient were more common in the PG (23.8% vs. 11.6% and 42.4% vs. 26.4%, respectively; p < 0.001 for both). The time on MV after the failure in the first SBT was higher in the PG than in the NPG (9 ± 5 days vs. 7 ± 2 days; p < 0.001). Conclusions: In this sample of difficult-to-wean patients, the use of a weaning protocol improved the decision-making process, decreasing the possibility of extubation failure.

Keywords: Ventilator weaning; Ventilation; Ventilators, mechanical.