Glycemic index explained: what it is, food chart and whether it matters for weight loss

What the glycemic index is and how it is measured

The concept was proposed in 1981 by David Jenkins and his team at the University of Toronto: they gave 62 foods to healthy fasting volunteers, measured their glucose for 2 hours and expressed the area under the curve as a percentage of the one produced by the same amount of carbohydrate taken as glucose. The current method, standardised in ISO 26642, is the same: 10 or more healthy people eat a portion with 50 g of available carbohydrate (total minus fibre), glucose is measured for 2 hours, and the area of the rise above the fasting level is divided by the one caused by 50 g of glucose, then multiplied by 100. A nuance almost nobody explains: GI does not measure the "speed" of the rise but its total size over those 2 hours. The cut-offs in the international tables are:

• Low GI: 55 or less (legumes, most fruits, dairy, pasta)
• Medium GI: 56 to 69 (brown rice, couscous, honey)
• High GI: 70 or more (white bread, boiled potato, white rice, corn flakes)

Glycemic load: the number that accounts for the portion

GI always compares 50 g of carbohydrate, but nobody eats "50 g of watermelon carbohydrate": you eat a serving. Glycemic load (GL) corrects for that with a simple sum: GL = GI × grams of available carbohydrate in the serving / 100. The usual cut-offs are: low, 10 or less; medium, 11 to 19; high, 20 or more. Two examples using the figures in the table below: 150 g of watermelon provide 10.7 g of available carbohydrate, so its load is 76 × 10.7 / 100 = 8 (low) despite its high GI. By contrast, 180 g of cooked spaghetti provide 52.3 g, and its load is 49 × 52.3 / 100 = 26 (high) despite its low GI. Pasta is not "bad", but the portion counts for more than its index.

Chart: glycemic index and load of 18 common foods

The GI values come from the summary table of the international tables by Atkinson, Foster-Powell and Brand-Miller (Diabetes Care, 2008), which average studies from several laboratories; the 2021 edition expanded the catalogue to more than 4,000 foods. We calculated the glycemic load from the available carbohydrate in each serving according to USDA FoodData Central. Treat them as a guide: the GI of the same food changes with the variety, ripeness, brand and cooking method. If you want to know how much carbohydrate and how many calories your real portion has, check the food calorie table.

What the chart teaches

• "Wholemeal" does not guarantee a low GI: wholemeal wheat bread (74) and white bread (75) are equally high; wholemeal wins on fibre, not on index
• Processing matters: traditional rolled oats as porridge have a GI of 55 and instant oats 79; instant mashed potato reaches 87 versus 78 for boiled potato
• Legumes are the champions: lentils 32 and chickpeas 28, with loads of 6 and 8 per plate
• Low GI with a big portion can mean a high load: spaghetti (GI 49) reaches 26 on a 180 g plate

What changes the glycemic index of a food

GI is not a fixed property like calories. The international tables themselves warn that bread, breakfast cereals, rice and snacks come in high-, medium- and low-GI versions, and that there are low-GI potato varieties even though most are high. What moves it most:

• Processing and cooking: the more pre-cooked and gelatinised the starch (instant flakes, flaked mash), the higher the GI tends to be
• Variety: there are rices and potatoes with clearly different GIs within the same type
• The fat and protein on the same plate: as early as 1981 Jenkins observed that the more fat and protein a food had, the smaller the glucose rise
• The person: the same meal does not produce the same curve in everyone, as the 800-person study from the Weizmann Institute showed

What really flattens your spikes

• Combine the carb with protein and/or fat (rice with chicken and oil, not rice alone)
• Add fibre: vegetables, legumes, whole fruit
• Eat the vegetables and protein first and the carb last: in a pilot study with 11 people with type 2 diabetes, eating vegetables and chicken 15 minutes before the bread and juice left glucose 37% lower at 60 minutes than the reverse order, with the very same meal
• Avoid liquid carbs (juices, soft drinks): a can of cola adds a load of 21 and does not fill you up
• Move after eating: a meta-analysis of 7 trials found that breaking up sitting time with gentle walks lowers glucose after meals, and more than simply standing

What the studies say: diabetes, weight and long-term health

This is where the most exaggeration happens, so here are the meta-analysis figures, which are modest:

• Diabetes: a Cochrane review of 11 trials (402 people) found HbA1c about 0.5 percentage points lower with low-GI diets; the largest meta-analysis (BMJ, 2021: 29 comparisons, 1,617 people with type 1 and type 2 diabetes) estimated 0.31 points less, with high certainty, plus small improvements in fasting glucose, LDL cholesterol and weight. Its authors call it a small but important improvement, on top of medication
• Weight: the 2007 Cochrane review (6 trials, 202 people with overweight, from 5 weeks to 6 months) found 1.1 kg more weight loss with low GI; however, a meta-analysis of 14 long-term trials found no differences in weight or body measurements, although it did find lower fasting insulin and lower C-reactive protein
• Long-term health: in 10 large cohorts of more than 100,000 participants each, people eating with the highest GI had a 27% higher risk of type 2 diabetes and a 15% higher risk of cardiovascular disease than those with the lowest. These are observational data, and the authors themselves note that the associations resemble those of eating more fibre and whole grains
• Compared with fibre: the series of systematic reviews partly funded by the WHO (Lancet, 2019) rated the evidence on GI and glycemic load as low or very low certainty, versus moderate evidence for fibre

Who does it really matter for?

The 2015 international expert consensus considers GI a valid and reproducible method, particularly relevant with insulin resistance, and asks for it always to be used within a diet that is already healthy, alongside fibre and whole grains. In plain terms: with diabetes, prediabetes or insulin resistance, choosing lower-load carbs is a tool with a measurable benefit (more context in what to eat with diabetes or prediabetes and what to eat with insulin resistance; specific adjustments belong with your healthcare team). If you are healthy, it is not worth obsessing over: with minimally processed food, enough protein and fibre and suitable calories, your glycemic response will already be good. And glucose rising after a meal is not a disease: it is how the body works.

Instead of memorising indexes, photograph your plates with Renzy: the AI estimates the macros and gives you a score that rewards balanced meals, which are exactly the ones that keep your glucose stable.