Quick answer: The glycemic index (GI) measures how quickly a single food raises blood glucose compared to pure glucose, scored 0 to 100. The glycemic load (GL) extends the idea by accounting for portion size — a more useful real-world number. Low GI doesn’t automatically mean “healthy” (peanuts and ice cream are low GI), and high GI doesn’t automatically mean “unhealthy” (watermelon and boiled potatoes are high GI). For most people without diabetes, GI is one of several factors to consider, not a single answer to what to eat.
What the glycemic index actually measures
The glycemic index was developed at the University of Toronto in the early 1980s by David Jenkins and colleagues. The protocol is specific: a healthy person eats a portion of a single test food containing 50 grams of available carbohydrate, and their blood glucose is measured at intervals over the next two hours. The area under the resulting blood-glucose curve is compared to the same person’s response to 50 grams of pure glucose (or, in some protocols, to white bread). The ratio, expressed as a percentage, is the food’s GI.
By convention:
- Low GI: 55 or less
- Medium GI: 56 to 69
- High GI: 70 or more
Some examples (the published GI tables vary slightly between studies):
- Glucose: 100 (the reference)
- White bread: 70 to 75
- Boiled white potato: 78
- Boiled brown rice: 50 to 68
- Porridge oats (cooked): 55
- Banana (ripe): 51 to 62
- Apple: 36
- Lentils (cooked): 32
- Whole milk: 39
- Peanuts: 14
Notice the range: peanuts and lentils are very low GI; boiled potatoes are very high. This is sometimes counter-intuitive — boiled potatoes are usually thought of as a healthy whole food, while ice cream is sometimes around 50 (low GI) because the fat content slows gastric emptying.
Why glycemic load is the more useful number
The GI uses a fixed 50 g carbohydrate test portion. In practice, you don’t eat 50 g of carbohydrate from every food — you eat a portion. And portion size matters. Watermelon has a high GI (around 76) but a normal serving contains so little total carbohydrate that the actual blood-glucose response is small.
Glycemic load adjusts for this. The formula is:
GL = (GI × grams of available carbohydrate per portion) ÷ 100
Conventional cut-offs:
- Low GL: 10 or less per portion
- Medium GL: 11 to 19
- High GL: 20 or more
Worked examples:
- 120 g of watermelon: high GI (76), low carbohydrate (~9 g) → GL of about 7. Low impact.
- 120 g of boiled potato: high GI (78), moderate carbohydrate (~21 g) → GL of about 16. Medium impact.
- 40 g of cornflakes with 125 ml semi-skimmed milk: high GI cereal (around 80), moderate carbohydrate (~32 g from cereal alone) → GL well above 20. High impact.
For most practical purposes — meal-planning, portion-thinking, comparing breakfast options — GL is the more useful number than GI alone.
Where GI / GL genuinely matters clinically
Type 2 diabetes management. NICE NG28 (recommendation 1.3.3) advises “eating high-fibre, low-glycaemic-index sources of carbohydrate, such as fruit, vegetables, wholegrains and pulses” as part of dietary advice, alongside medication, weight management, and physical activity. Beyond NG28 itself, meta-analyses (Cochrane 2009; Livesey et al. 2019) report modest but consistent HbA1c reductions when low-GI choices are made over months.
Type 1 diabetes carbohydrate counting. People using insulin to manage type 1 diabetes count carbohydrates by gram for meal-time insulin dosing. GI / GL helps fine-tune the timing and quantity of insulin, but the count of total carbohydrate is the primary number.
Gestational diabetes. NHS dietary guidance for gestational diabetes uses similar principles — slower-releasing carbohydrates spread across the day.
Reactive hypoglycaemia / dumping syndrome. A small subset of people experience symptomatic blood-sugar drops 1 to 3 hours after high-GI meals; reducing GI helps.
Where GI / GL is much less useful
- For people without diabetes who eat mixed meals. When carbohydrate is consumed alongside protein, fat, and fibre (a normal meal), the GI of any individual ingredient becomes a poor predictor of the blood-glucose response. The mixed-meal effect dominates.
- For weight management. Total energy intake remains the primary driver of weight change. GI / GL has not been shown to consistently outperform total-calorie approaches in randomised trials.
- For “healthiness” judgements. Peanuts have a GI of 14 because of their fat content; ice cream is around 50 for the same reason; chocolate cake can be in the medium range. Low GI does not equal healthy. The Eatwell Guide reasoning (variety, fibre, vegetables, water, modest free sugars) does the consumer-facing job better.
What changes a food’s GI
- Ripeness. A green banana has a GI around 30; a fully ripe banana around 60. Resistant starch in the green fruit converts to free sugars during ripening.
- Cooking and processing. The longer a starchy food is cooked or the more it is processed (mashing, blending, finely milling), the higher its GI tends to climb. Al dente pasta is lower GI than overcooked pasta. Steel-cut oats are lower GI than instant oats.
- Cooling and reheating. Cooked-then-cooled rice and potato develop resistant starch — the GI drops. Reheating doesn’t fully reverse this. The “leftover rice” effect is real, though modest.
- What it’s eaten with. Fat, protein, and fibre alongside a carbohydrate slow gastric emptying and absorption — the meal-as-a-whole has a lower GI than the carbohydrate component alone.
- Order of eating. Eating vegetables and protein before the carbohydrate portion of a meal blunts the post-meal glucose rise (small but consistent in trials in people with type 2 diabetes).
Practical takeaways
- If you don’t have diabetes, the GI / GL of individual foods is not a daily decision-making tool. The Eatwell Guide is enough — variety of vegetables, plenty of fibre, modest free sugars.
- If you have type 2 diabetes, low-GI carbohydrate choices (porridge oats over cornflakes; lentils alongside rice; whole fruit over juice) are evidence-supported, alongside the rest of NICE NG28 guidance.
- If you use a continuous glucose monitor (CGM), you’ll see your own personal responses to specific foods. Use those — your individual response can vary substantially from published GI tables.
- Don’t use low GI as a heuristic for “healthy”. Pair it with the broader food question: vegetables, fibre, protein quality, free-sugar limit.
The bottom line
The glycemic index is a useful tool for managing type 2 diabetes and for understanding how single carbohydrate foods affect blood glucose. The glycemic load extends GI by accounting for portion size — and is more relevant for everyday food decisions. Both numbers are descriptive, not normative: low GI is not the same as healthy, and high GI is not the same as bad. For people without diabetes, the broader Eatwell-Guide framing of variety, fibre, vegetables, and modest free sugars does the consumer-facing job better than chasing GI numbers.
Related Camden Medicals reading
Sources and further reading
- NICE NG28 — Type 2 diabetes in adults: management
- Diabetes UK — Glycaemic index and diabetes
- SACN — Carbohydrates and Health (2015)
- NHS — Starchy foods and carbohydrates
Camden Medicals editorial. Information, not medical advice — if you have diabetes, are pregnant, or are managing a metabolic condition, talk to your GP, dietitian, or pharmacist about how to apply this guidance to your situation.