Abstract

Magnetic resonance imaging (MRI) was used to characterize the volumes and rates of gastric emptying of both liquid and gas following the ingestion of beverages of varying carbonation and carbohydrate levels. Eight subjects drank 800 ml each of four test beverages in a counterbalanced order: water, a non-carbonated carbohydrate-electrolyte solution (NC), a lightly carbonated carbohydrate-electrolyte solution (PC), and a carbonated cola (CC). T2-weighted, echoplanar images (25-30 continuous slices, 1 cm thick, 256 x 128 matrix. TE = 80, 40 cm FOV) of the abdomen were collected at minutes 3, 110, 20, 30, 45, and 60 following beverage ingestion. Images were analyzed for gas and liquid volumes. Water and NC emptied the most rapidly, with half times of 21 (3) and 31 (3) min, respectively [mean (SE)]. PC emptied significantly slower [47 (6) min] and CC slower yet [107 (8) min]. the carbonation content of the beverage accounted for 84% of the variation in emptying time, whereas carbohydrate content did not account for any significant variation. The gastric gas volume of the CC was higher at 2 min post-ingestion compared with all other drinks; however, the rate of emptying of the gas was the same among all beverages. Significantly greater total gastric volumes (gas + liquid) were associated with the ingestion of CC, and accordingly produced a greater severity of gastric distress, as evaluated with a gastric distress inventory. The high gastric gas volumes (=600ml) after ingestion of CC suggested a potential source of error in body composition using standard hydrostatic weighing methods. This prediction was tested in nine additional subjects. Ingestion of 800 ml of CC prior to hydrostatic weighing resulted in a 0.7% underestimate of body density and thus an 11% overestimate of percentage body fat compared to measurements made before beverage consumption.