OBJECTIVE To determine the fasted and fed gastrointestinal motility characteristics that are possibly responsible for gastric retention in mechanically ventilated patients. METHODS Prospective, case series. METHODS Surgical intensive care unit of a university hospital. METHODS Seven patients who required mechanical ventilation for thoracic or combined thoracic-neurologic injuries and nine healthy volunteers. METHODS None. RESULTS Antroduodenal manometry was performed during fasting and gastric feeding with a polymeric diet in patients during mechanical ventilation, weaning, and after detubation. Gastric retention volumes were determined during gastric tube feeding. Motility data were compared with recordings from nine healthy volunteers. During the fasting state, under sedation and morphine, the migrating motor complex in patients was significantly (p < .001) shortened: median 32.0 vs. 101.0 mins in healthy volunteers. During gastric tube feeding, the motility pattern did not convert to a normal postprandial pattern until morphine was discontinued. An interdigestive or mixed interdigestive-postprandial pattern was seen during gastric tube feeding in most patients during morphine administration. Most (94%) of the activity fronts during gastric feeding started in the duodenum. Gastric retention percentages during gastric tube feeding were negatively correlated (r2=.44; p < .01) with antral motor activity. CONCLUSIONS These data suggest that morphine administration affects antroduodenal motility in mechanically ventilated patients. The gastrointestinal motor pattern involved in impaired gastric emptying in morphine-treated patients is characterized by antral hypomotility and persisting duodenal activity fronts during continuous intragastric feeding. The observed motility patterns suggest that early administration of enteral feeding might be more effective into the duodenum or jejunum than into the stomach of mechanically ventilated patients.