Glycaemic indexGI

The glycaemic index ranks carbohydrate foods by how much they raise blood glucose in the two hours after eating, compared with a reference of pure glucose scored at 100. It was developed in the early 1980s as a research tool and became a diet framework afterwards.

Its main weakness is that it is measured per 50 g of available carbohydrate rather than per portion. Watermelon has a high index and so little carbohydrate per slice that eating it changes very little, which is why glycaemic load, index adjusted for the actual portion, is the more useful number.

It is also unstable in the ways real eating is. Ripeness, cooking method, cooling and reheating, processing, and above all what else is on the plate all move it. Add fat, protein or fibre and the same carbohydrate produces a flatter curve, so an index measured on a food eaten alone rarely describes a meal.

Individual variation is the other complication. Continuous glucose monitoring studies find people responding quite differently to identical meals, which undermines any universal table and is also the honest basis of the personalised nutrition industry.

Where it earns its place is in managing diabetes, where lower glycaemic index diets have modest evidence for improving glucose control. For people without diabetes the case is much weaker, and a post meal rise is normal physiology rather than an injury.

Most of what the index tracks in practice is fibre and processing, which are worth attending to directly. Whole foods with their fibre intact tend to produce gentler curves, and that is a simpler rule than a lookup table.

In practice, build meals from minimally processed carbohydrate with protein and fibre alongside, and treat the index as background rather than a shopping guide.