The Precision of in vitro Methods and Algorithms for Predicting the Bioavailability of Dietary Iron

Abstract

Three factors determine how much iron will be absorbed from a meal. They are the physiological mechanisms that regulate uptake by and transfer through the enterocytes in the upper small intestine, the quantity of iron in the meal, and its availability to the cellular iron transporters. Established methods exist for predicting the effect of physiological regulation and for measuring or estimating meal iron content. Three approaches to estimating bioavailability have been advocated. Two are in vitro screening procedures: measurement of dialyzable iron and Caco-2 cell uptake, both carried out after in vitro simulated gastric and pancreatic digestion. The third is the use of algorithms based on the predicted effects of specific meal components on absorption derived from isotopic studies in human volunteers. The in vitro procedures have been very useful for identifying and characterizing factors that affect non-heme iron absorption, but direct comparisons between absorption predicted from the in vitro tests and measurements in human volunteers have only been made in a limited number of published studies. The available data indicate that dialysis and Caco-2 cell uptake are useful for ranking meals and single food items in terms of predicted iron bioavailability, but may not reflect the magnitudes of the effects of factors that influence absorption accurately. Algorithms based on estimates of the amounts of heme iron and of enhancers and inhibitors of non-heme iron absorption in foods make it possible to classify meals or diets as being of high, medium, or low bioavailability. The precision with which meal iron bioavailability can be predicted in a population, for which a specific algorithm has been developed, is improved by measuring the content of the most important enhancers and inhibitors. However, the accuracy of such predictions appears to be much lower when the algorithm is applied to meals eaten by different populations.