RELIABILITY AND INTEROBSERVER VARIABILITY OF ULTRASONOGRAPHIC MEASUREMENT OF GASTRIC-EMPTYING RATE

Scintigraphy and real-time ultrasound are valid techniques to measure parameters of gastric emptying. However, scintigraphy involves exposur e to ionizing radiation, while the precision of real-time ultrasound h as not been previously evaluated. The objectives of the present study were to determine the inter observer and day-to-day variability of liq uid gastric emptying rate measured by real-time ultrasonography in hea lthy volunteers and to compare the gastric emptying rate of males and females. Twenty healthy volunteers had ultrasonographic measurement of gastric emptying rate after ingestion of 300 ml beef broth. Nine subj ects had a concurrent assessment by a second observer. Nine subjects h ad two studies performed on different days by the same sonographer. Th e T1/2 for 20 subjects was 24.77 +/- 6.84 min with no difference betwe en the half-emptying time for males (25.89 +/- 6.99 min) and females ( 24.02 +/- 6.94 min). The Pearson and intraclass correlation coefficien ts for observations made by two observers were 0.83 and 0.625, with a difference due to observer of 2.37 min +/- 5.26 (NS). The test-retest reliability across successive days was 0.136, representing considerabl e day-to-day variability within subjects. The variability between subj ects was also large, explaining up to 79% of the total variance. We co nclude that ultrasound is a useful method to evaluate gastric emptying with good interobserver agreement. Due to substantial day-to-day vari ability, sample sizes larger than previously suggested are required to demonstrate clinically important changes in gastric emptying rate in clinical trials.