The purpose of these studies was to develop a radiologic method for assessing gastric emptying of an indigestible solid in humans and to apply this technique to the evaluation of patients with diabetes mellitus. Thirty healthy subjects ingested 10 solid radiopaque markers (small pieces of nasogastric tubing) together with a standard meal (donuts and 7-Up). Radiographs of the upper abdomen were obtained hourly for up to 6 h until all markers had emptied from the stomach. Although most of the liquid component of the meal, labeled with 111In, emptied during the first hour (as assessed simultaneously by radionuclide scintigraphy), few radiopaque markers emptied from the stomach during the first 2 h after the meal. Most markers emptied during the fourth postprandial hour, and all 10 markers had emptied by 6 h in 45 of 46 experiments. In contrast, not all of the solid radiopaque markers emptied from the stomach by 6 h in 16 of 26 experiments in patients with diabetes mellitus (p less than 0.001 vs. healthy controls). In some experiments, 99mTc-labeled scrambled eggs were added to the meal so that emptying of this digestible solid, assessed by scintigraphy, could be compared with emptying of liquids and solid radiopaque markers. In healthy subjects, the digestible solid emptied more slowly than the liquid (t 1/2 = 154 +/- 11 min vs. 30 +/- 3 min, p less than 0.001), but emptying of digestible solid was significantly faster than the emptying of the indigestible solid radiopaque markers. In diabetics, emptying rates for the digestible solid and liquid were close to normal (t 1/2 = 178 +/- 5 min and 40 +/- 3 min, respectively), whereas indigestible solid markers were retained in the stomach 6 h after the meal in 50% of the patients. Radiopaque markers proved to be a simple method for measuring gastric emptying of indigestible solids in humans. Using this technique, patients with insulin-dependent diabetes mellitus had a high incidence of abnormally slow gastric emptying of indigestible solids; the method may be a more sensitive indicator of gastric motor dysfunction than radionuclide scintigraphy.