Objective: It is assumed that delayed gastric emptying (GE) occurs frequently in critical illness; however, the prevalence of slow GE has not previously been assessed using scintigraphy. Furthermore, breath tests could potentially provide a convenient method of quantifying GE, but have not been validated in this setting. The aims of this study were to (i) determine the prevalence of delayed GE in unselected, critically ill patients and (ii) evaluate the relationships between GE as measured by scintigraphy and carbon breath test. Design: Prospective observational study. Setting: Mixed medical/surgical intensive care unit. Patients: 25 unselected, mechanically ventilated patients (age 66 years (49-72); and 14 healthy subjects (age 62 years (19-84)). Interventions: GE was measured using scintigraphy and 14C-breath test. A test meal of 100 ml Ensure (standard liquid feed) labelled with 14C octanoic acid and 99mTechnetium sulphur colloid was placed in the stomach via a nasogastric tube. Main outcome measures: Gastric 'meal' retention (scintigraphy) at 60, 120, 180 and 240 min, breath test t 50 (BTt 50), and GE coefficient were determined. Results: Of the 24 patients with scintigraphic data, GE was delayed at 120 min in 12 (50%). Breath tests correlated well with scintigraphy in both patients and healthy subjects (% retention at 120 min vs BTt 50; r 2=0.57 healthy; r 2=0.56 patients; p≤0.002 for both). Conclusions: GE of liquid nutrient is delayed in approximately 50% of critically ill patients. Breath tests correlate well with scintigraphy and are a valid method of GE measurement in this group.