To develop a statistical model for predicting zinc bioavailability from cereal-based vegetarian meals using relative proportion of nutrients, non-nutrients and their interactive effects.
A database on in vitro zinc dialysability (by isotopic tracer, 65Zn) of vegetarian meals (266 out of 326) from Asia, Africa, Europe /US and Latin America was used to develop a model for estimating zinc bioavailability. A multiple regression analysis adjusted for energy content was carried out for net bioavailable zinc from a meal with the predictor variables as meal contents of iron, zinc, copper, ascorbic acid, beta-carotene, riboflavin, thiamine, folic acid, tannic acid, fiber, phytate degradation products (IP6 to IP1), along with their interaction terms. Reproducibility of the model was tested with remaining 60 meals. Validation of the model was done with zinc absorption data of i) 12 young adults on 24 meals and ii) 5 adults with ileostomy on 7 meals.
Folic acid, IP3 and IP5 were significant influencing factors for bioavailable zinc. Weighted multiple regression equation was: ln (bioavailable zinc in mg) = -1.701 + 1.285 x ln [(IP5 in mg] x (Zn in mg)] -1.222 x ln(IP5 in mg) -0.0078 x folic acid in microg -0.137 x ln [(IP3 in mg) x (Zn in mg)] with adjusted R-- = 0.64, p = 0.0001. The correlation between predicted and observed dialysability of meals was found to be 0.96 (p < 0.01). A significant correlation between observed and predicted amount of absorbed zinc (r = 0.85, p < 0.01) was obtained for the human data of zinc absorption in 12 healthy and 5 subjects with ileostomy.
Bioavailable amount of zinc from vegetarian meals was influenced by IP3, IP5 and folic acid content and their interactive effect with zinc content.