Criteria for performing a spontaneous breathing trial (SBT) have not been evaluated in controlled trials. An important component of these criteria is neurological status. The objective of this study was to evaluate whether physicians take mental status into consideration before performing an SBT in mechanically ventilated patients. This was a prospective, observational study which included 355 mechanically ventilated patients. Daily assessments were made of whether the patients met criteria for performing a SBT. On the day a patient met the criteria, the level of sedation was evaluated using the Glasgow Coma Scale as modified by Cook and Palma (GCS-Cook) and it was registered whether or not the physician carried out an SBT. Two hundred and four patients (57%) underwent an SBT on the day they met the criteria (cohort 1) and in 151 patients (cohort 2) the SBT was delayed a median time of 1 day (interquartile range 1–2). There were differences in the GCS-Cook score on the day the criteria were met for performing an SBT (mean 13 ± 3 points in cohort 1 versus 9 ± 3 points in cohort 2; P < 0.001). There were differences ( P < 0.001) between the cohorts in days of intubation and length of stay in the intensive care unit. Neurological status/level of sedation is a factor in the decision whether or not to perform a spontaneous breathing trial.