Although the WIND (Weaning according to a New Definition) classification based on duration of ventilation after the first separation attempt has been proposed, this new classification has not been tested in clinical practice. The objective of this cohort study was to evaluate the clinical relevance of WIND classification and its association with hospital mortality compared to the International Consensus Conference (ICC) classification.

All consecutive medical ICU patients who were mechanically ventilated for more than 24 h between July 2010 and September 2013 were prospectively registered. Patients were classified into simple, difficult, or prolonged weaning group according to ICC classification and Groups 1, 2, 3, or no weaning (NW) according to WIND classification.

During the study period, a total of 1600 patients were eligible. These patients were classified by the WIND classification as follows: Group NW = 580 (36.3%), Group 1 = 617 (38.6%), Group 2 = 186 (11.6%), and Group 3 = 217 (13.6%). However, only 735 (45.9%) patients were classified by ICC classification as follows: simple weaning = 503 (68.4%), difficult weaning = 145 (19.7%), and prolonged weaning = 87 (11.8%). Clinical outcomes were significantly different across weaning groups by ICC classification and WIND classification. However, there were no statistical differences in successful weaning rate (96.6% vs. 95.2%) or hospital mortality (22.5% vs. 25.5%) between simple and difficult weaning groups by the ICC. Conversely, there were statistically significant differences in successful weaning rate (98.5% vs. 76.9%) and hospital mortality (21.2% vs. 33.9%) between Group 1 and Group 2 by WIND.

The WIND classification could be a better tool for predicting weaning outcomes than the ICC classification.