In vitro starch digestion kinetics, corrected for estimated gastric emptying, predict portal glucose appearance in pigs.

In vitro starch digestion is used for predicting the in vivo glucose response, but their relationship has not been defined thoroughly. To clarify, in vitro starch digestion using a modified Englyst-assay was compared to portal glucose appearance in pigs. Four portal vein-catheterized pigs (43.2 +/- 4.8 kg body weight) were fed 4 diets containing 70% purified starch ranging from slowly to rapidly digestible [maximal rate of in vitro digestion (%)/min: 0.22 (slowly), 0.38, 0.73, and 1.06 (rapidly)] for 7-d periods in a 4 x 4 Latin square. In vivo (R2 = 0.964) and in vitro (R2 = 0.998) data were modeled using a Chapman-Richards model that accurately described the sigmoidal glucose-release profiles. Across samples, the extent of glucose recovered was less in vivo than in vitro (69 vs. 42% of starch). The rate of glucose release adjusted for plateau effects was lower in vivo (0.35 vs. 0.89%/min), whereas the shape parameter adjusted for plateau effects (sigmoidal modifier) was higher in vivo (37.9 vs. 13.7). Consequently, peak glucose release in vivo occurred 69 min postprandial, whereas it occurred only 6 min into the second stage of digestion in vitro. Cumulative portal glucose appearance was strongly related (R2 = 0.89; P < 0.001) to in vitro glucose release, although a nonlinear bias was observed. After correcting in vitro release with predicted gastric emptying, the relationship improved and became linear (R2 = 0.95; P < 0.001). In conclusion, in vitro starch digestion kinetics predict portal glucose appearance up to 8 h postprandial accurately provided that in vitro data are corrected for gastric emptying.