Designing Clean-Label Low-Glycemic Foods: From Starch Structure to Human Glycemic Response

Clean-label strategies to lower the glycemic impact of starch-based foods are attracting growing interest, but supporting evidence remains fragmented across starch science, food engineering, digestion methodology, and clinical nutrition. This narrative review, informed by a structured literature search, integrates this evidence into a bottom–up framework linking raw material selection, physical processing, food matrix interactions, in vitro digestion, and human glycemic testing. Using pigmented rice as a model system, it examines how heat–moisture treatment, annealing, ultrasound, extrusion, microwave-assisted processing, and controlled retrogradation reorganize starch, and how hydrocolloids, proteins, lipids, and phenolic compounds reduce enzyme accessibility or enzyme activity. This review makes three contributions. First, mechanistic claims are graded as directly demonstrated, inferred from digestibility data, or hypothetical, and D-allulose–protein conjugation is explicitly treated as a hypothesis. Second, Englyst fractions, INFOGEST kinetics, hydrolysis index, estimated GI, and ISO 26642 human GI are compared as non-interchangeable outputs. Third, an implementation table specifies, for each tier, the design variable, analytical method, evidence required, and principal limitation. The framework is a decision support tool rather than a validated predictive model; formal low-GI claims require human testing, and applications such as 3D-printed foods for older adults require separate validation.