The objective of this work was to identify the unique characteristics of patients who experienced success in a 30-min spontaneous breathing trial (SBT) but failed at 120 min. Patients who had received mechanical ventilation for >24 h were eligible for inclusion in this study. The SBT was performed by 7 cm H2O of pressure support with zero PEEP. After a successful 120-min SBT, weaning from mechanical ventilation was performed. Data were collected at 30 and 120 min or at the failure of the SBT. All patients who successfully completed a 30-min SBT were enrolled. We enrolled 352 subjects in this study. Of these, 311 subjects (88.4%) directly completed a 120-min SBT (success group), and 41 subjects (11.6%) passed the test for at least 30 min but failed before 120 min (failure group). In data collected before the SBT, presence of chronic cardiopulmonary disease, number of previous SBT attempts before 30-min SBT success, age, and PaCO2 were independently associated with 120-min SBT failure. A scale was developed that used these 4 variables. The failure rate was low in subjects with ≤2 points (3%) but significantly higher in subjects with >2 points (46%). In data collected at the 30-min SBT, PaCO2 , rapid shallow breathing index, ΔPaO2 /FIO2 , Δbreathing frequency, and ΔpH were independently associated with 120-min SBT failure. These 5 variables were then used to develop another scale to predict SBT success. Similar to the previous score, the failure rate was low in subjects with ≤2 points (1%) and significantly higher in subjects with >2 points (55%). This study highlights differences between subjects who completed a 120-min SBT and those who succeeded at 30 min but failed by 120 min. In subjects with a score >2 points, reflecting a greater risk of SBT failure, a 120-min SBT may be required. Copyright © 2018 by Daedalus Enterprises.