Data from clinical studies established that there was an inverse linear relationship between measures of postprandial blood glucose and insulin responses to an oral glucose load, consumed in a drink, and the logarithm of viscosity of the drink. These data have been re-analysed using concentration and molecular weight as the dependent variables. Molecular weight (M) of the beta-glucans used was determined using high-performance size exclusion chromatography equipped with a triple detector system of right angle light scattering, viscometry and refractive index. A significant relationship between changes in peak blood glucose and a combination of logarithm of the concentration and logarithm of M was found.