The glycaemic index of a food is not a physical constant. When researchers in Boston measured the GI of white bread in 63 healthy adults, using the standard protocol and three repeats per person, they got a mean of 62 with a standard deviation of 15. The same person, tested three times on the same bread, varied by 20% from one occasion to the next. Between people, the spread reached 25%. Adding participants did not help. Extending the blood sampling did not help either.

That is the first thing to know before building a menu around this number. The second is more useful: in the largest synthesis ever published on carbohydrate quality, dietary fibre and whole grains came out with a substantially higher grade of evidence than glycaemic index for predicting type 2 diabetes, cardiovascular disease and mortality. The most popular marker is the weakest of the three.

This article covers where GI came from, what it actually measures, what the large studies do and do not find, and what to build a plant-based diet on when you want steady blood sugar. With numbers, portions and the sources.

What the glycaemic index actually measures

The glycaemic index was born in 1981, in a paper by David Jenkins and colleagues in the American Journal of Clinical Nutrition. The protocol: 62 common foods fed individually to groups of 5 to 10 healthy fasting volunteers. Blood glucose is followed for two hours, and the resulting area under the curve is expressed as a percentage of the area produced by the same amount of carbohydrate taken as pure glucose.

Three details of that protocol matter enormously and almost always vanish in popular write-ups.

The test uses 50 g of available carbohydrate, not a portion. To test watermelon you have to eat roughly 700 g of it. To test white bread, two and a half slices will do. Comparing the GI of those two foods therefore means comparing two situations nobody lives in.

The food is eaten alone, fasted. No mixed meal, no coffee, no olive oil, no exercise in the preceding hours.

The sample is tiny. Five to ten people per food in the original study, ten in the current ISO method. It is an order of magnitude, not a precision measurement.

The usual thresholds, as published by Diabetes UK, are simple: low GI up to 55, medium from 56 to 69, high from 70. What Diabetes UK also spells out, and what tends to get dropped: GI takes no account of portion size, which limits its practical use from the start.

What Jenkins had already found in 1981

The original 1981 table is worth rereading, because it contradicts a common shortcut. By category, the rises in blood glucose ran as follows: dried legumes 31%, dairy products 35%, fruit 50%, cereals and biscuits 60%, breakfast cereals 65%, vegetables 70%. Legumes topped the ranking from the start, and they have stayed there.

But the paper also noted a significant negative relationship between the glucose rise and fat content (p < 0.01) and protein content (p < 0.001), and no significant relationship with fibre content. From the founding publication onwards, fat lowered GI more reliably than fibre did. That is why a doughnut can post a lower GI than a steamed potato, which ought to be enough on its own to disqualify GI as a sole selection criterion.

How stable is the number?

The study by Nirupa Matthan and colleagues, published in the American Journal of Clinical Nutrition in 2016, is the most explicit on this. Sixty-three adults free of chronic disease, aged 18 to 85, with BMI from 20 to 35. Each underwent three sets of food challenges, reference glucose and white bread, both at 50 g of available carbohydrate, with glucose and insulin followed for five hours.

The result: white bread GI of 62 ± 15. Intra-individual coefficient of variation 20%, inter-individual 25%. The authors then tested whether precision could be improved by increasing sample size, replicating both reference and test foods, lengthening blood sampling and changing the area-under-the-curve calculation. None of it reduced the coefficients of variation. Their conclusion is blunt: it is unlikely that GI is a good approach for guiding food choices.

Two large continuous glucose monitoring studies confirmed the problem from another angle. Zeevi and colleagues, in Cell in 2015, followed 800 people for a week and measured responses to 46,898 meals, finding high variability in the response to identical meals. The PREDICT 1 study, published in Nature Medicine in 2020 on 1,002 UK adults, quantified that spread: after strictly identical meals, the coefficient of variation between participants reached 68% for glucose, 59% for insulin and 103% for triglycerides.

PREDICT 1 adds an important nuance that often gets distorted. For blood glucose specifically, the meal’s macronutrient composition explained 15.4% of the variance, against 6.0% for person-specific factors. What you eat does matter, and it matters more than your individual profile for the glycaemic response. The problem with GI is not that the food is irrelevant. It is that the number attached to that food is too imprecise to choose between two close options.

What the large syntheses find

Two levels of evidence answer two different questions, and they need to be kept apart.

At population level

Cohort studies follow tens of thousands of people for years and record who develops diabetes. A dose-response meta-analysis by Geoffrey Livesey and colleagues, published in Nutrients in 2019, pooled the studies judged methodologically valid. The relative risk of type 2 diabetes was 1.27 (95% confidence interval 1.15 to 1.40) per 10 GI units, across 10 studies, and 1.26 (1.15 to 1.37) per 80 g/day of glycaemic load in a 2,000 kcal diet, across 15 studies. Over the full range of observed intakes, from 47.6 to 76.1 GI units, the relative risk reached 1.87 (1.56 to 2.25).

Those figures are real and the association is robust. But they are associations. People whose average dietary GI is high also typically eat fewer pulses, fewer vegetables and more ultra-processed food.

At trial level

Randomised trials speak to causation, and here the picture gets considerably less tidy.

A meta-analysis published in Current Diabetes Reviews in 2026 pooled 25 randomised controlled trials and 1,973 participants aged 12 to 75. Low-GI diets reduced HbA1c by 0.22 percentage points (p < 0.01), fasting blood glucose by 0.19 mmol/L (p = 0.04) and LDL cholesterol by 0.15 mmol/L (p < 0.01). Effects on HDL, fasting insulin, HOMA-IR, triglycerides and total cholesterol did not reach statistical significance. Benefits were stronger in people with type 2 diabetes, in those whose fasting glucose exceeded 6.1 mmol/L and in trials lasting at least twelve weeks.

A real effect, then, but concentrated in people whose blood sugar is already disturbed, and a modest one: 0.22 points of HbA1c is roughly a quarter of what first-line medication achieves.

The trial that gets left out

OmniCarb, published in JAMA in 2014 by Frank Sacks and colleagues, is the most rigorous trial ever run on the question and the one enthusiastic articles cite least. One hundred and sixty-three overweight adults, four complete diets of five weeks each, with every meal, snack and calorie-containing drink supplied by the study, all built on a DASH-type base.

At high carbohydrate content, moving from GI 65 to GI 40 reduced insulin sensitivity from 8.9 to 7.1 units, a fall of 20% (p = 0.002), and raised LDL cholesterol from 139 to 147 mg/dL, a rise of 6% (p ≤ 0.001). No effect on HDL, triglycerides or blood pressure. That is the opposite of what you usually read, and it comes from the best controlled study in the field.

The verdict from the broadest synthesis

In 2019, Andrew Reynolds and colleagues published in The Lancet a series of systematic reviews commissioned by the WHO, covering 185 prospective studies and 58 clinical trials, close to 135 million person-years of data. The aim was to compare the various markers of carbohydrate quality against each other.

The fibre results are clear: a 15 to 30% reduction in all-cause and cardiovascular mortality, and in the incidence of coronary heart disease, stroke, type 2 diabetes and colorectal cancer, comparing the highest fibre consumers with the lowest. Benefit was greatest between 25 and 29 g of fibre a day, and the dose-response curves suggested further benefit above that. Whole grains produced comparable findings.

For glycaemic index and glycaemic load, the authors report smaller or absent risk reductions. And crucially, they grade the certainty of the evidence: moderate for dietary fibre, low to moderate for whole grains, low to very low for glycaemic index and glycaemic load.

That grading is the basis on which the WHO published its updated carbohydrate guidance on 17 July 2023. It sets no target GI. It states that carbohydrate intake for everyone aged 2 and over should come primarily from whole grains, vegetables, fruits and pulses, and it sets thresholds: at least 400 g of vegetables and fruits and at least 25 g of naturally occurring dietary fibre per day for adults. For children, at least 250 g of fruit and vegetables and 15 g of fibre from 2 to 5 years, 350 g and 21 g from 6 to 9 years, 400 g and 25 g from 10 years.

The NHS sets the adult fibre target at 30 g a day and notes that average intake sits around 20 g. Among 11 to 18 year olds it is 16 g against a 25 g target. The shortfall is real and it has been counted.

Glycaemic load, with the formula

Glycaemic load fixes the main flaw in GI, which is that it ignores quantity. The formula fits on one line:

GL = (GI × grams of available carbohydrate in the portion) / 100

Three calculations are enough to see what that changes.

Watermelon has a GI around 72, which is high. A 150 g portion supplies roughly 11 g of carbohydrate. Glycaemic load: 72 × 11 / 100 = 8. That is low.

A 45 g portion of rolled oats, the lower end of the Swiss FSVO portion, supplies about 27 g of carbohydrate at a GI around 55. Glycaemic load: 15.

A 250 g portion of boiled potatoes, within the FSVO range, supplies about 40 g of carbohydrate at a GI often near 80. Glycaemic load: 32.

Watermelon, a high-GI food, therefore produces a load four times lower than rolled oats, a low-GI food. The usual thresholds put a low load below 10, medium at 11 to 19 and high from 20 upwards, per portion.

Glycaemic load is more informative than GI, but it inherits the same imprecision, because it is calculated from the same unstable number. It is useful for avoiding gross errors, not for choosing between two options that are close together.

Plant foods, ranked honestly

The international tables published by Fiona Atkinson, Jennie Brand-Miller and colleagues in the American Journal of Clinical Nutrition in 2021 list over 4,000 items, a 61% increase on the 2008 edition. They are split into two lists: around 2,100 values derived using the ISO-recommended methodology, and around 1,900 determined by less robust methods, with limited numbers of subjects or a large standard error. That split says everything: half the values in circulation are not reliable.

What the tables do establish solidly:

Legumes, pasta, fruits and dairy products are consistently low-GI, with values that hold up across countries. It is the most stable finding of the last forty years.

Cereals and cereal products, including whole-grain and wholemeal versions, show wide variation instead, which the authors attribute to differences in manufacturing methods. Breads, breakfast cereals, rice and savoury snacks exist in low, medium and high-GI versions. “Wholemeal” is therefore not a synonym for “low GI”.

Most potato varieties are high-GI, but specific low-GI varieties have now been identified.

Whole grain or wholemeal, the distinction that counts

Diabetes UK puts this distinction better than most articles. A wholegrain food contains grains that remain intact, forming a physical barrier that slows enzyme access to the starch. A wholemeal food contains the whole grain, but ground up. The result: some mixed-grain breads with visible grains have a lower GI than finely milled wholemeal bread, and sometimes than white bread.

In practice, for a flatter glucose curve, a bread with grains you can see and feel beats a wholemeal loaf with an even crumb at comparable fibre content. Structure does the work, not composition alone.

The levers that genuinely change the response

Five levers are documented by human data. They are worth more than a GI table.

Starch structure and cooling

Work published in Foods in 2026 measured the effect of post-cooking storage on boiled rice. GI fell from 83.03 ± 15.02 on freshly prepared rice to 43.55 ± 6.99 after prolonged freezing, while resistant starch rose from roughly 1.8% to nearly 4.0%. The correlation between resistant starch and GI was strong and negative (r = -0.935, p < 0.001).

The mechanism is retrogradation: as it cools, amylose reorganises into crystalline structures that amylases digest poorly. The fraction withdrawn from digestion behaves like a fibre and feeds the colonic microbiota.

In practice: cook rice, pasta or potatoes the day before, chill them in the fridge, then reheat. Retrogradation partly survives reheating. Note the standard deviations, though: ± 15 on fresh rice means the effect is real but your rice will not land exactly at 43.

Cooking texture

A randomised trial published in Nutrients in 2026 compared hard-cooked and soft-cooked potatoes, each replacing one third of the carbohydrate from rice in a meal. The hard-cooked version retained more resistant starch and more total phenolics, required more chews, a longer oral sensory exposure time and a longer eating duration. It significantly reduced the glucose incremental area under the curve, peak glucose and glycaemic excursion, along with peak insulin and the insulin resistance index.

The effect did not carry over to the following meal. But within the meal itself, a cooking method that leaves some resistance to the bite beats a soft one at identical ingredient. Al dente pasta works through the same mechanism.

Beta-glucans, at the right dose

This is the only mechanism on the list with full regulatory backing. In an opinion published in the EFSA Journal in 2026, the EFSA NDA Panel concludes that a cause-and-effect relationship has been established between consumption of oat beta-glucans and reduction of postprandial blood glucose peaks, based on 16 human intervention studies. The condition is precise: the food or meal must contain at least 30 g of available carbohydrate per portion and at least 3 g of oat beta-glucans per 30 g of available carbohydrate.

A meta-regression published in the European Journal of Clinical Nutrition in 2023, covering 59 comparisons and 340 participants, adds that the minimum effective dose depends heavily on the molecular weight of the beta-glucans: 0.2 g per 30 g of available carbohydrate at high molecular weight (above 1,000 kg/mol), 2.2 g at medium weight (300 to 1,000 kg/mol) and 3.2 g at low weight (below 300 kg/mol).

That explains a frustrating phenomenon: industrial processing, prolonged cooking and fine milling all reduce beta-glucan molecular weight. A porridge made from jumbo oats and cooked briefly, and an industrial oat drink, can contain the same quantity of beta-glucans and not produce the same effect. Our piece on the real cost of homemade versus industrial oat milk takes that question further.

Legumes, with the usual caveats

Legumes have been the most consistently low-GI food since 1981, and the Swiss FSVO recommends eating them at least once a week, with a portion of 60 g dry weight.

On acute effects, a meta-analysis published in Nutrition Journal in 2025 pooled 28 studies and 40 comparisons on chickpeas. Postprandial glucose incremental area under the curve was significantly reduced against carbohydrate-matched controls (mean difference -47.89, 95% CI -64.20 to -31.58, p < 0.0001). But there was no significant effect on peak glucose or on the insulin response, and the GRADE certainty of evidence is rated very low because of heterogeneity.

On medium-term effects, a systematic review published in Nutrients in 2020 examined 18 randomised trials lasting at least six weeks. Only the five trials in people with type 2 diabetes reported significant effects, on fasting glucose, HbA1c, fasting insulin or 2-hour postprandial glucose. Overall certainty of evidence there is also rated very low.

Legumes remain the best available choice on a plant-based plate, for their fibre, protein and nutrient density. But the promise of a metabolic transformation from adding lentils alone does not rest on solid evidence. Our guide to legume batch cooking, from soaking to freezing and the one on the benefits of soaking and sprouting cover the practical side.

What else is on the plate

Because GI is measured on a single food eaten fasted, everything served alongside changes the outcome. Jenkins measured this back in 1981: fat and protein lowered the glucose response to a statistically significant degree, fibre did not.

The corollary is uncomfortable: fat lowers GI. A doughnut, a crisp or an ice cream can post a lower GI than a steamed potato or a slice of watermelon. Using GI as a food quality criterion therefore produces absurd rankings by construction. Diabetes UK states it plainly: fat lowers the GI of a food.

Potatoes: what the cohorts actually say

Potatoes are the plant food most systematically condemned on GI grounds. The long-term data tell a more interesting story.

A study published in the BMJ in 2025 followed 205,107 participants across three US cohorts over 5,175,501 person-years, with 22,299 diagnoses of type 2 diabetes. Per increment of three servings a week:

• potatoes of all kinds: 5% higher risk (HR 1.05, 95% CI 1.02 to 1.08);
• French fries: 20% higher risk (HR 1.20, 95% CI 1.12 to 1.28);
• baked, boiled or mashed: HR 1.01 (0.98 to 1.05), meaning no significant association.

In substitution analyses, replacing three weekly servings of potatoes with whole grains lowered the risk by 8% for total potatoes, 4% for the non-fried versions and 19% for chips. Replacing potatoes with white rice raised the risk. A meta-analysis of 13 cohorts, 587,081 participants and 43,471 cases confirmed the pattern.

GI does not distinguish a boiled potato from a chip. The long-term data do, and they flip the practical conclusion entirely. Our article on getting the most out of potatoes in batch cooking covers the cooking methods.

Putting it into practice, with numbers

The Swiss FSVO nutrition recommendations give the most directly usable framework in Switzerland.

Grain products and potatoes: 3 portions a day, at least half of them wholegrain. One portion is 75 to 125 g of bread, or 200 to 300 g of potatoes, or 45 to 75 g of flakes, pasta, rice, semolina, buckwheat, quinoa or flour, dry weight.

Fruit and vegetables: 5 portions a day, meaning 3 of vegetables and 2 of fruit.

Legumes: at least once a week, 60 g dry weight per portion, within the daily protein-rich portion.

Seeds and nuts: a small handful a day, 15 to 30 g.

On that basis, here is a plant-based day built on the markers that hold up rather than on GI values quoted to the decimal:

Morning. 50 g of jumbo oats, cooked briefly to preserve beta-glucan molecular weight, with 20 g of pumpkin seeds and 100 g of berries. Roughly 8 to 10 g of fibre.

Midday. 200 g of cooked lentils, from 70 g dry, over wholegrains cooked the day before and chilled, with 200 g of raw vegetables and a tablespoon of rapeseed oil. Roughly 15 g of fibre.

Snack. A whole apple with 20 g of almonds. Whole fruit rather than juice, because the fruit’s structure slows access to its sugars.

Evening. Wholemeal pasta cooked al dente, chickpea and vegetable sauce, a portion of seeds. Roughly 10 g of fibre.

Day total: between 33 and 38 g of fibre, above the NHS target of 30 g and well above the WHO minimum of 25 g. No GI value was consulted to build that menu. Our pages on combining grains and legumes effectively and whole grains in your diet go into the pairings.

Six common mistakes

Treating GI values as constants. A table says “quinoa 53”. That number probably comes from a test on ten people with a coefficient of variation around 20%. The real interval sits closer to 42 to 64.

Comparing foods on GI without looking at portions. Glycaemic load exists for this reason. Watermelon at GI 72 gives a load of 8; oats at GI 55 give a load of 15.

Assuming “wholemeal” means “low GI”. The 2021 international tables are explicit: cereals, wholegrain versions included, show the widest variation of any category. Some finely milled wholemeal breads are high-GI.

Chasing the lowest possible GI. In OmniCarb, at high carbohydrate content, the low-GI diet produced a 20% fall in insulin sensitivity and a 6% rise in LDL. Lowering GI is not a goal in itself.

Avoiding potatoes because of their GI while eating white rice. The BMJ cohorts show the substitution runs the wrong way: swapping non-fried potatoes for white rice raises diabetes risk.

Counting on fibre to flatten an acute glucose spike. On the short-term response, the 1981 study found no significant relationship with fibre while it did find one for fat and protein. The major benefit of fibre is chronic, measured over years, and it is solidly established by the 2019 Lancet series. That is a different time horizon.

Adapting to different situations

Type 2 diabetes. This is the group with the most favourable data. The 25-trial meta-analysis reports stronger effects when fasting glucose exceeds 6.1 mmol/L, and the five legume trials in people with type 2 diabetes are the only ones in that body of work to show significant effects. Any dietary change should be discussed with the care team, particularly if glucose-lowering medication is in use, because changing the meal can change the requirement.

Type 1 diabetes. Carbohydrate counting and insulin adjustment remain the foundation. GI can help anticipate the kinetics, but with 20% intra-individual variability it replaces no measurement-guided adjustment. This belongs in a specialist consultation.

Pregnancy and gestational diabetes. Follow-up is medically supervised and protocols differ by country. The general FSVO and WHO framework applies; implementation belongs with the obstetric team.

Endurance athletes. The goal partly inverts: before and during exercise, rapid absorption is the point. The glucose rise is not a problem there but a function. See vegetarian diet and endurance sports.

Sensitive guts. Jumping from 20 to 35 g of fibre a day causes bloating and discomfort. The NHS recommends increasing gradually over several weeks, with adequate fluid. Soaked, well-cooked legumes are better tolerated.

Children. The WHO puts numbers on it: at least 250 g of fruit and vegetables and 15 g of fibre from 2 to 5 years, 350 g and 21 g from 6 to 9, 400 g and 25 g from 10. Building a child’s menu around GI values has no basis.

Limits of the available data

Three caveats deserve to be stated plainly.

The GI values quoted here, even those from the 2021 international tables, remain orders of magnitude. Around 1,900 of the 4,000 entries in those tables come from methods the authors themselves describe as less robust.

The cohort data on potatoes, GI and diabetes are associations. Statistical substitution is not the same as real substitution in people’s lives, and those who eat a lot of chips differ from everyone else in many other ways.

The conclusions on starch retrogradation and cooking texture rest on acute, small trials measuring postprandial responses. None has shown that a long-term health benefit follows. It is a plausible, well-documented mechanism, not a treatment.

Frequently asked questions

Is the glycaemic index useless? No, but it is oversold. It remains useful for spotting gross differences, such as lentils against a puffed rice cake. It is unusable for choosing between two foods whose values differ by less than 15 points, because that gap is smaller than the measurement error.

Why do official recommendations not mention it? Because the evidence does not support it. The 2019 Lancet series grades certainty as low to very low for GI and glycaemic load, against moderate for fibre. The WHO’s 2023 carbohydrate recommendation is built on foods and on fibre, with no GI target.

Does cooling my rice really lower its glycaemic index? On the boiled rice tested in Foods in 2026, measured GI fell from 83 to 44 after prolonged freezing, with resistant starch roughly doubling. The effect is real and the mechanism is established. The exact size of it for your rice, with your cooling time, is uncertain.

How much beta-glucan do I need for an effect on the glucose peak? EFSA has established the cause-and-effect relationship and sets the condition at a minimum of 3 g of oat beta-glucans per 30 g of available carbohydrate. The 2023 meta-regression shows molecular weight matters as much as dose: 0.2 g is enough at high molecular weight, 3.2 g is needed at low.

Does fruit raise blood sugar? The international tables place fruit among the categories that are consistently low-GI, with values that hold across countries. The WHO recommends at least 400 g of fruit and vegetables a day for adults. Whole fruit beats juice, because the fruit’s structure slows access to the sugars.

Is wholemeal bread always better than white? On fibre, yes. On GI, not necessarily: Diabetes UK lists some wholemeal breads among high-GI foods, and a bread with visible whole grains does better than finely milled wholemeal. Structure beats milling.

Do insulin spikes make you gain weight? This is the most widespread and least supported claim in the field. OmniCarb, with every meal supplied for five weeks per diet, did not find the expected metabolic benefits from the low-GI diet, and even measured a drop in insulin sensitivity at high carbohydrate content. On this specific point, the intuitive mechanics do not survive controlled trials.

For the neighbouring measure, see understanding the insulin index of plant foods, and on sweeteners, natural sugar alternatives and GI. For the practical side of grains, see the cooking guide for barley, spelt and oats and the top 10 essential legumes for your pantry.

Sources

• Jenkins DJA et al., “Glycemic index of foods: a physiological basis for carbohydrate exchange”, American Journal of Clinical Nutrition, 1981. Europe PMC
• Atkinson FS et al., “International tables of glycemic index and glycemic load values 2021”, American Journal of Clinical Nutrition, 2021. Europe PMC
• Matthan NR et al., “Estimating the reliability of glycemic index values and potential sources of methodological and biological variability”, American Journal of Clinical Nutrition, 2016. Europe PMC
• Berry SE et al., “Human postprandial responses to food and potential for precision nutrition” (PREDICT 1), Nature Medicine, 2020. Europe PMC
• Zeevi D et al., “Personalized Nutrition by Prediction of Glycemic Responses”, Cell, 2015. Europe PMC
• Livesey G et al., “Dietary Glycemic Index and Load and the Risk of Type 2 Diabetes: A Systematic Review and Updated Meta-Analyses of Prospective Cohort Studies”, Nutrients, 2019. Europe PMC
• “Low Glycemic Index Diets on Glycemic and Lipid Control Across Diverse Populations: A Systematic Review and Meta-Analysis”, Current Diabetes Reviews, 2026. Europe PMC
• Sacks FM et al., “Effects of high vs low glycemic index of dietary carbohydrate on cardiovascular disease risk factors and insulin sensitivity: the OmniCarb randomized clinical trial”, JAMA, 2014. Europe PMC
• Reynolds A et al., “Carbohydrate quality and human health: a series of systematic reviews and meta-analyses”, The Lancet, 2019. Europe PMC
• “Total and specific potato intake and risk of type 2 diabetes”, BMJ, 2025. Europe PMC
• “The Effects of Legume Consumption on Markers of Glycaemic Control”, Nutrients, 2020. Europe PMC
• “Chickpea attenuates postprandial blood glucose responses: a systematic review and meta-analysis”, Nutrition Journal, 2025. Europe PMC
• EFSA Panel on Nutrition, Novel Foods and Food Allergens, “Oat beta-glucans and reduction of postprandial glucose peak”, EFSA Journal, 2026. Europe PMC
• “The importance of molecular weight in determining the minimum dose of oat beta-glucan required to reduce the glycaemic response”, European Journal of Clinical Nutrition, 2023. Europe PMC
• “Impact of Post-Cooking Storage on the Glycemic Profile of Boiled Rice”, Foods, 2026. Europe PMC
• “Minimally Cooked Potato Improved Glycemic Response Across Two Meals”, Nutrients, 2026. Europe PMC
• World Health Organization, “WHO updates guidelines on fats and carbohydrates”, 17 July 2023. who.int
• NHS, “How to get more fibre into your diet”. nhs.uk
• Diabetes UK, “Glycaemic index and diabetes”. diabetes.org.uk
• FSVO, “Swiss nutrition recommendations for adults”. blv.admin.ch

This article is informational and does not replace individual medical advice. If you have diabetes, are pregnant or are on medication, dietary decisions belong with a healthcare professional.