Why the physical structure of a food changes what it does to you, independent of the nutrients it contains.
What the evidence says
- The core proposition: food, not the nutrient, is the fundamental unit of nutrition. Whole foods exert effects greater than what their individual nutrient components predict — “food synergy” — because the physical structure of the food determines how those nutrients are digested and absorbed[1]
- A reductionist focus on isolated nutrients has produced misleading dietary advice and misleading product design[1]
- The cleanest experimental separation of structure from composition. 58 adults received preloads matched for weight (266 g), energy (125 kcal), energy density, and ingestion rate, in five conditions: whole apple, applesauce, apple juice with added fiber, apple juice without fiber, and no preload[2]
- Subsequent meal intake as a percentage of control: whole apple 85%, applesauce 94%, juice with fiber ~99%, juice without fiber ~103% — a graded effect tracking how intact the structure was[2]
- Eating a whole apple before the meal cut total energy intake by 15% (187 kcal)[2]
- Adding fiber back to juice produced no satiety benefit — no significant difference between juice with and without fiber, and the added fiber reduced palatability[2]
- That last result is the load-bearing one, and the bullet above names why it cannot carry as much as it looks: the fiber-added juice was also less palatable, so the arm differed in taste as well as fiber. It weakens fiber content as the sole explanation and points toward structural properties — volume, chewing, intact cell walls — without ruling fiber out[2]
- Degree of processing tracks the effect across a wide food set — in a study its authors call preliminary, whose unit of analysis is foods rather than people. Across 98 ready-to-eat foods classified by NOVA, the degree of processing correlated strongly with both higher glycemic response and lower satiety potential. Notably, traditional glycemic index did not correlate with processing degree, while glycemic glucose equivalent (which accounts for actual portions) did[3]
- A measured mechanism for the speed difference: removing fiber from a food accelerated gastric emptying 186.0 (15.6) versus 231.7 (17.3) minutes, the paper’s own values and dispersion, and hunger returned more rapidly after the low-fiber meal[4]
- Fiber within an intact structure slows gastric emptying, blunts glycemic response, promotes satiation, and alters gut hormone secretion — the glycaemic figure, with what the paper actually reports: area under the glycaemic curve ”226 (23.1) v 160 (20.0) mmol/min/dl, p<0.05", low- versus high-fibre meal. `csi05` gives two values and no percentage; "29% lower" is arithmetic on them, not a reported result — and it is 29% only against the low-fibre value, so write the two numbers, or say which way round the 29% runs[4],[5]