Influence of moist-heat processing and intrinsic food matrix characteristics on iron bioaccessibility in underutilized green leafy vegetables: a stable isotope-assisted INFOGEST 2.0 assessment.

Underutilized green leafy vegetables (GLVs) are promising plant-based iron sources, yet their nutritional potential is limited by poor iron bioaccessibility. This study investigated iron bioaccessibility in Amaranthus tricolor L., Amaranthus viridis L., and Brassica oleracea var. botrytis using a modified INFOGEST 2.0 in vitro digestion model incorporating stable isotope-assisted (57Fe) quantification. Proximate composition, mineral profile, antinutritional factors, intrinsic food matrix characteristics, and the effects of common moist-heat processing were evaluated to identify the determinants of iron bioaccessibility, revealing distinct species-specific differences in the food matrix architecture, chemical functionality, and phenolic composition. Moist-heat processing significantly (p < 0.05) reduced phytate and oxalate contents by up to 78% and 83%, respectively, while increasing the iron solubility from 10.35-16.58% in unprocessed samples to 18.35-26.59% after processing and the iron dialyzability from 4.05-4.42% to 6.60-14.92%. Although boiling produced the greatest proportional improvement in iron bioaccessibility, blanching provided a more favorable balance between iron retention and bioaccessibility and yielded the highest absolute dialyzable iron concentrations across the three GLVs. Iron dialyzability was strongly and negatively correlated with oxalate (r = -0.833, p < 0.001) and phytate (r = -0.732, p = 0.007) and strongly positively correlated with iron solubility (r = 0.858, p < 0.001), indicating that total iron content alone did not predict the proportion of iron rendered bioaccessible during digestion. These findings demonstrate that iron bioaccessibility is determined by the combined influence of intrinsic food matrix characteristics, processing-induced modifications, and reductions in antinutritional factors, providing mechanistic insight into why total iron content alone is a poor predictor of nutritional potential.