A food-group based algorithm to predict non-heme iron absorption.

Objective: To develop an algorithm to predict the percentage non-heme iron absorption based on the foods contained in a meal (wholemeal cereal, tea, cheese, etc.). Existing algorithms use food constituents (phytate, polyphenols, calcium, etc.), which can be difficult to obtain.

Design: A meta-analysis of published studies using erythrocyte incorporation of radio-isotopic iron to measure non-heme iron absorption.

Methods: A database was compiled and foods were categorized into food groups likely to modify non-heme iron absorption. Absorption data were then adjusted to a common iron status and a weighted multiple regression was performed.

Results: Data from 53 research papers (3,942 individual meals) were used to produce an algorithm to predict non-heme iron absorption (R(2) =0.22, P < 0.0001).

Conclusions: The percentage non-heme iron absorption can be predicted from information on the types of foods contained in a meal with similar efficacy to that of food-constituent-based algorithms (R(2) = 0.16, P= 0.0001).