Story
Glycemic Index & Glycemic Load
Last updated
In one pass In the small intestine, starch is broken down into glucose, absorbed into the blood, and blood sugar rises.
Educational content, not medical advice — consult a clinician.
Story path
Chapter 1
GI is a lab curve
In the small intestine, starch is broken down into glucose, absorbed into the blood, and blood sugar rises. The glycemic index () compares how much blood sugar rises in total over 2 hours when a food delivers a set amount of available carbohydrate.
The test fixes just one thing. Healthy volunteers who have fasted eat a portion of the food containing 50 g of available carbohydrate (roughly total carbohydrate minus fiber). The area under their blood-glucose curve over the next 2 hours is recorded and compared with the same 50 g of pure glucose, which is set at 100. Jenkins and colleagues proposed the method in 1981.
A high GI means only this: at that standard portion, the food raises blood sugar faster and further. It is an index of a food — not the post-meal glucose peak itself, not a health score, and certainly not your own curve. Each test group is usually only 5 to 10 people, and the GI is their average. So when you see a GI number, remember the three conditions behind it: a fixed 50 g, a fasted person, and the average of a small group.
The test fixes just one thing. Healthy volunteers who have fasted eat a portion of the food containing 50 g of available carbohydrate (roughly total carbohydrate minus fiber). The area under their blood-glucose curve over the next 2 hours is recorded and compared with the same 50 g of pure glucose, which is set at 100. Jenkins and colleagues proposed the method in 1981.
A high GI means only this: at that standard portion, the food raises blood sugar faster and further. It is an index of a food — not the post-meal glucose peak itself, not a health score, and certainly not your own curve. Each test group is usually only 5 to 10 people, and the GI is their average. So when you see a GI number, remember the three conditions behind it: a fixed 50 g, a fasted person, and the average of a small group.
Background · Why 50 g, and why fiber is left out
Why fix 50 g of available carbohydrate? To compare fairly how fast different foods raise blood sugar under the same sugar load, you have to hold the amount of carbohydrate constant and let only the food's own structure (particle size, fiber, fat, processing) shape the curve.Why available carbohydrate, with fiber subtracted? Human digestive enzymes cannot break fiber apart. It barely raises blood sugar directly and instead travels to the colon to feed gut bacteria (see Carbs & Fiber). What actually becomes glucose in the small intestine and enters the blood is the starch and sugar.
Why the area under the 2-hour curve rather than the highest point? A sharp peak with a fast fall and a gentle low peak with a long tail can share the same peak value yet mean different things for the body. The area counts both how high and how long.
So a high is not a verdict on whether to eat a food. It is an index averaged over a small group in a lab, at a fixed 50 g of available carbohydrate.
Chapter 2
Glycemic load adds the portion
The has a blind spot: it fixes the amount of carbohydrate at 50 g and ignores how much you actually eat. The total glucose that reaches your blood depends on how much available carbohydrate the meal contains. The measure that supplies this missing half is the glycemic load ().
The formula: GL = GI × the serving's actual available carbohydrate (grams) ÷ 100. Salmerón and colleagues introduced the idea in 1997, first in a cohort study of women and then in one of men.
So avoiding every high-GI food across the board is crude. A meal's glucose exposure is the GI multiplied by the portion you eat. Watermelon is often treated as a high-GI food, but a slice holds little carbohydrate that can raise blood sugar, so its GL is low; a big bowl of rice has a high GL.
The formula: GL = GI × the serving's actual available carbohydrate (grams) ÷ 100. Salmerón and colleagues introduced the idea in 1997, first in a cohort study of women and then in one of men.
So avoiding every high-GI food across the board is crude. A meal's glucose exposure is the GI multiplied by the portion you eat. Watermelon is often treated as a high-GI food, but a slice holds little carbohydrate that can raise blood sugar, so its GL is low; a big bowl of rice has a high GL.
Numbers · Watermelon, rice, and GI values that drift
Treating watermelon as a food with a high glycemic index () is an old textbook example. A normal slice (about 120 g of flesh) holds very little carbohydrate that can raise blood sugar, so one serving has a of only about 4 (low). A bowl of white rice (GI about 73) reaches a GL of about 30 per serving (high); lentils (GI about 30) come to a GL of about 8 per serving (low).One thing is worth saying plainly: the GI of the same food drifts a lot from study to study. Early single measurements put watermelon at 72 to 76 (high), while the 2021 international tables of GI and GL values, which pooled thousands of data points from around the world, revised its average GI down to about 50 (low to moderate).
That is not a contradiction. It confirms what the GI is: a measured group average, noisy by nature, swayed by variety, ripeness, the test group, and how blood was sampled after the meal. Memorizing a food's GI to the last digit and treating it as law misreads the measure.
Only two conclusions are solid. Multiplying in the portion (GL) is closer to reality than the GI alone. And what else is in the meal, and who you are, matter more than the number in the table.
Chapter 3
Absorption speed shapes the spike
What makes a food's glycemic index () high or low actually happens in the small intestine and the pancreas.
In the small intestine, enzymes split starch and sugar into glucose, which crosses the gut wall into the blood. The faster the splitting and absorption, the sharper and higher the post-meal glucose peak; the slower, the lower and flatter the curve. Beta cells in the pancreas sense the rise and release insulin to move the sugar into cells, and the steeper the rise, the bigger that wave of insulin.
So the GI is an outside readout of how fast glucose is released in the small intestine. Fat and protein in the same meal, the viscosity of fiber, particle size, and cooling after cooking (when starch recrystallizes) all slow that release. Eating rice after it has cooled did lower the post-meal rise a little in one small trial, but the effect was modest — far smaller than switching to intact grains and beans.
In the small intestine, enzymes split starch and sugar into glucose, which crosses the gut wall into the blood. The faster the splitting and absorption, the sharper and higher the post-meal glucose peak; the slower, the lower and flatter the curve. Beta cells in the pancreas sense the rise and release insulin to move the sugar into cells, and the steeper the rise, the bigger that wave of insulin.
So the GI is an outside readout of how fast glucose is released in the small intestine. Fat and protein in the same meal, the viscosity of fiber, particle size, and cooling after cooking (when starch recrystallizes) all slow that release. Eating rice after it has cooled did lower the post-meal rise a little in one small trial, but the effect was modest — far smaller than switching to intact grains and beans.
Evidence · cooled rice really does blunt the rise
Anything that slows glucose release in the small intestine lowers the post-meal glucose peak from the same portion of carbohydrate:Fat and protein in the same meal slow stomach emptying and digestion. Jenkins's 1981 experiments already saw that the more fat and protein a food carried, the smaller the post-meal rise in blood sugar.Fiber and viscosity: soluble fibers such as oat beta-glucan and those in beans thicken the gut contents, so glucose reaches the gut wall and is absorbed more slowly.Processing and particle size: the finer the milling and the softer the cooking, the easier it is for enzymes to reach the starch, and the faster blood sugar rises (the gap between intact grain and flour).Resistant starch from cooling: when cooked rice or potatoes cool, part of the starch recrystallizes into resistant starch that the body digests poorly, and the blood-sugar response falls.
Cooling-and-recrystallizing is not folklore. A 2015 study measured white rice prepared three ways: freshly cooked, cooled at room temperature for 10 hours, and refrigerated at 4°C for 24 hours and then reheated. In the refrigerated-and-reheated rice, resistant starch rose from 0.64 to 1.65 g per 100 g; in the randomized crossover trial that followed, 15 healthy adults ate that rice and freshly cooked rice, and the post-meal glucose response was statistically significantly lower after the first.
To be honest about the effect size: this was a small trial of 15 people, and the difference was real but modest. It did not turn white rice into a low- staple. What it confirms is the mechanism: the physical structure of starch (not just whether it is a carbohydrate) sets how fast the small intestine can break it down. To get more of this effect, switch to intact-grain, bean, and high-fiber carbohydrates rather than fussing over the temperature of your rice.
Chapter 4
Same food, different people
The is an average across a group. The same banana sends one person's blood sugar soaring and barely moves another's.
The differences come from a string of personal factors: gut bacteria, everyday insulin sensitivity, how well you slept, what you ate at the previous meal, whether you just exercised, and the time of day. Fat, protein, and fiber in the same meal, plus who you are, together make the GI table far less absolute.
That is not a reason to throw the table away; just do not mistake the table for your body. Do not swing to the other extreme either. In people without diabetes, a brief rise in blood sugar after a meal that comes back down is the body working normally, and there is no need to hunt down every small peak.
The differences come from a string of personal factors: gut bacteria, everyday insulin sensitivity, how well you slept, what you ate at the previous meal, whether you just exercised, and the time of day. Fat, protein, and fiber in the same meal, plus who you are, together make the GI table far less absolute.
That is not a reason to throw the table away; just do not mistake the table for your body. Do not swing to the other extreme either. In people without diabetes, a brief rise in blood sugar after a meal that comes back down is the body working normally, and there is no need to hunt down every small peak.
Myth · Should healthy people fight every glucose spike?
The evidence that exposes the average comes from putting glucose sensors on many people and having them eat the same foods. In 2015, the Weizmann Institute in Israel (Zeevi and colleagues, published in Cell) fitted 800 people with continuous glucose monitors (, small sensors worn on the skin that read blood sugar around the clock) and recorded nearly 47,000 real post-meal glucose responses. The finding: for the same food, the differences in glucose response between people were larger than the differences between foods. One person's blood sugar shot up after a banana but stayed flat after cookies; another person showed the reverse. A model the researchers trained on personal features, including gut bacteria, predicted a given person's response to a given meal fairly accurately.As CGMs went mainstream they brought a side effect: the anxiety that "any rise in blood sugar is bad." Time to tap the brakes.
In healthy people, blood sugar is supposed to rise after a meal. That is the body working normally, not a fault. A brief physiological swing up and back down after eating is not the same thing as the persistently high blood sugar of diabetes. For people without diabetes, treating every small glucose peak as a threat to be hunted down has no hard evidence behind it, and it slides easily into over-restricting food and food anxiety.
Where a CGM genuinely helps is much narrower: people who already have prediabetes or type 2 diabetes, or who have repeated symptoms after meals, can wear one for a week or two to find their own few trigger foods (see CGM for the Metabolically Healthy). For a healthy person watching blood sugar out of pure curiosity, the payoff is limited and the cost is anxiety.
Chapter 5
Low GI is no weight-loss shortcut
Low is often sold as a magic switch for weight loss and metabolism. Lay the evidence out and it is far less impressive, and where the evidence is strong is not where the hype is.
The relatively solid part is blood-sugar management in type 2 diabetes: a Cochrane systematic review pooling randomized trials found that low-GI diets modestly lowered (HbA1c, a lab test that reflects average blood sugar over the past few months). The most hyped and least supported part is healthy people losing weight through low GI: the Cochrane review updated in 2023 judged the evidence insufficient to draw conclusions. The main line of weight loss is still calories, and whether you can stick with a diet.
Do not slide the evidence from diabetes blood-sugar management over into proof of weight loss in healthy people.
The relatively solid part is blood-sugar management in type 2 diabetes: a Cochrane systematic review pooling randomized trials found that low-GI diets modestly lowered (HbA1c, a lab test that reflects average blood sugar over the past few months). The most hyped and least supported part is healthy people losing weight through low GI: the Cochrane review updated in 2023 judged the evidence insufficient to draw conclusions. The main line of weight loss is still calories, and whether you can stick with a diet.
Do not slide the evidence from diabetes blood-sugar management over into proof of weight loss in healthy people.
Evidence · Does the low-GI weight-loss claim hold?
The relatively solid part: blood-sugar management in diabetes. A 2009 Cochrane systematic review pooled 11 with 402 people with type 1 or type 2 diabetes and found that low- diets lowered the long-term blood-sugar marker (HbA1c) by about 0.5 percentage points on average. The effect was small but real, and it came without more low-blood-sugar episodes. This is low GI's most defensible use.Even this part is contested. The 2019 nutrition consensus report of the American Diabetes Association (ADA) notes that the two systematic reviews of the glycemic index (GI) and glycemic load () it cites found no significant effect on HbA1c in people with or at risk of diabetes, and mixed results on fasting glucose, and that studies define high and low GI in so many different ways that the clinical usefulness of GI and GL is uncertain. Useful, then, but not the decisive lever.
The most hyped and least supported part: healthy people losing weight through low GI. The Cochrane review updated in 2023 (10 studies, 1,210 people) concluded that the evidence is not enough to draw any firm conclusion, with certainty rated moderate to very low. The 2018 DIETFITS randomized trial (609 adults with overweight, followed for 12 months) was blunter: the healthy low-fat and healthy low-carbohydrate groups lost similar amounts of weight, and a person's insulin secretion did not predict which diet suited them better. That result does not support the claim that "carbohydrate and insulin are the master switch for weight loss."
The two claims, each set against the evidence:
Claim: low-GI foods automatically burn fat and lower weight. The 2023 Cochrane review says the evidence is not enough to conclude this; in DIETFITS, where both groups were coached to eat high-quality foods, the healthy low-fat and healthy low-carbohydrate groups lost similar amounts of weight, and insulin secretion did not predict which diet suited whom, so weight loss did not follow the single axis of carbohydrate and insulin.Claim: low GI is the key to reversing metabolism. In people with type 2 diabetes, low GI modestly improves HbA1c (by about 0.5 percentage points) and is worth using as a supporting tool; but the reviews the ADA cites found no significant HbA1c effect and the definitions are messy, so it is not a factor that decides things on its own.
Where does the gap come from? One phrase, "low GI," gets moved from the setting of blood-sugar management in diabetes to the setting of weight loss in healthy people. The first has controlled trials behind it; the second rests mostly on short, small studies with wobbly conclusions. Treating the first as proof of the second is the most common switch.
To be honest: if you have diabetes or prediabetes, low GI and low GL are a handle worth discussing with your doctor or dietitian. If you are healthy and simply want to lose weight, what decides success is overall diet quality, calories, and whether you can stick with it — not staring at a GI table.
Chapter 6
A guide, not a rule
The glycemic index () and glycemic load () are a lens for understanding food, not a list of good and evil.
Look at the carbohydrate quality of the whole plate, not at single numbers. Use pairing and food structure to lower the post-meal glucose peak, rather than giving up carbohydrates. If you have diabetes or prediabetes, they are a handle to use together with your doctor and dietitian; if you are healthy, one normal portion of a high-GI food is nothing to panic over.
This page is health education, not a diagnosis, and it does not replace individualized advice.
Look at the carbohydrate quality of the whole plate, not at single numbers. Use pairing and food structure to lower the post-meal glucose peak, rather than giving up carbohydrates. If you have diabetes or prediabetes, they are a handle to use together with your doctor and dietitian; if you are healthy, one normal portion of a high-GI food is nothing to panic over.
This page is health education, not a diagnosis, and it does not replace individualized advice.
In practice · Judge the plate's carbohydrate quality
A series of systematic reviews and , Reynolds 2019, puts carbohydrate quality — especially fiber — at the center of long-term health outcomes, and the Dietary Guidelines for Americans 2020-2025 likewise put the overall eating pattern ahead of any single number. Both point to the same thing: what governs long-term metabolic health is the carbohydrate quality of the whole plate — more intact grains, beans, and fiber-rich fruit and vegetables, less refined sugar and refined flour — not scoring every food by its glycemic index ().A sugary drink with a low GI (sweetened with fructose) does not become healthy because its GI is low. The GI was never a health score; it is only how fast 50 g of available carbohydrate raises blood sugar in a lab.
So the first use is not memorizing a table. It is asking one question: is the carbohydrate in this meal intact and fiber-bearing, or milled and sugared?
In practice · Pair foods, and apply it to you
The same portion of carbohydrate, eaten with vegetables, protein, and healthy fats — or eaten after the vegetables and protein — lowers the post-meal glucose peak by slowing absorption in the small intestine. Intact beats milled, and chewy beats cooked to mush. That is not giving up carbohydrates; it is changing how fast they reach the blood.Apply it to your own situation. If you have type 2 diabetes or prediabetes, the glycemic index () and glycemic load () are a handle to use together with your doctor and dietitian; the ADA's 2019 nutrition consensus puts the overall eating pattern and carbohydrate quality first, and GI and GL are not factors that decide things on their own. If you are healthy, one normal portion of a high-GI food inside a balanced meal is nothing to panic over.
Nutrition is a long game. This page is health education, not a diagnosis, and it does not replace individualized advice from a doctor or dietitian. Once you understand how glucose is released in the small intestine and why everyone differs, you are much harder to lead around by marketing that says "low GI cures everything."
References · 13
- Jenkins, D. J., Wolever, T. M., Taylor, R. H., Barker, H., Fielden, H., Baldwin, J. M., et al. (1981). Glycemic index of foods: a physiological basis for carbohydrate exchange. The American Journal of Clinical Nutrition, 34(3), 362-366. Foundational paper: 62 foods fed as 50 g available carbohydrate to groups of 5-10 healthy fasting volunteers, 2-hour glucose area under the curve expressed relative to glucose (=100); fat and protein were inversely related to the postprandial rise. 10.1093/ajcn/34.3.362
- Institute of Medicine. (2005). Dietary Reference Intakes for Energy, Carbohydrate, Fiber, Fat, Fatty Acids, Cholesterol, Protein, and Amino Acids. National Academies Press. nap.nationalacademies.org/catalog/10490/dietary-reference-intakes-for-energy-carbohydrate-fiber-fat-fatty-acids-cholesterol-protein-and-amino-acids
- Salmerón, J., Ascherio, A., Rimm, E. B., Colditz, G. A., Spiegelman, D., Jenkins, D. J., et al. (1997). Dietary fiber, glycemic load, and risk of NIDDM in men. Diabetes Care, 20(4), 545-550. Prospective cohort of 42,759 men; introduced dietary glycemic load (GI × available carbohydrate) as an exposure metric and found high glycemic load with low cereal fiber associated with higher NIDDM risk. 10.2337/diacare.20.4.545
- Salmeron, J., Manson, J. E., Stampfer, M. J., Colditz, G. A., Wing, A. L., & Willett, W. C. (1997). Dietary fiber, glycemic load, and risk of non-insulin-dependent diabetes mellitus in women. JAMA, 277(6), 472-477. High GL diets associated with increased T2DM risk in women (RR 1.47, highest vs lowest quintile; RR 2.50 for high GL combined with low cereal fiber; the 2.17 figure belongs to the companion men's paper); cereal fiber protective; GI/GL more predictive than total carbs. 10.1001/jama.1997.03540300040031
- Atkinson, F. S., Brand-Miller, J. C., Foster-Powell, K., Buyken, A. E., & Goletzke, J. (2021). International tables of glycemic index and glycemic load values 2021: a systematic review. The American Journal of Clinical Nutrition, 114(5), 1625-1632. Systematic compilation of >4000 GI/GL values; the pooled average GI of watermelon is ~50, illustrating that published GI values for a food vary widely between measurements. 10.1093/ajcn/nqab233
- Sonia, S., Witjaksono, F., & Ridwan, R. (2015). Effect of cooling of cooked white rice on resistant starch content and glycemic response. Asia Pacific Journal of Clinical Nutrition, 24(4), 620-625. Randomized single-blind crossover in 15 healthy adults; refrigerating cooked white rice at 4°C for 24 h then reheating raised resistant starch from 0.64 to 1.65 g/100 g and significantly lowered the postprandial glycemic response vs freshly cooked rice. pubmed.ncbi.nlm.nih.gov/26693746
- Zeevi, D., Korem, T., Zmora, N., Israeli, D., Rothschild, D., Weinberger, A., et al. (2015). Personalized nutrition by prediction of glycemic responses. Cell, 163(5), 1079-1094. 800 participants, ~46,898 postprandial glycemic responses measured by CGM; person-to-person variability to identical meals exceeded food-to-food variability, and a microbiome-inclusive model predicted individual responses. 10.1016/j.cell.2015.11.001
- Evert, A. B., Dennison, M., Gardner, C. D., Garvey, W. T., Lau, K. H. K., MacLeod, J., et al. (2019). Nutrition therapy for adults with diabetes or prediabetes: a consensus report. Diabetes Care, 42(5), 731-754. ADA consensus notes that systematic reviews of GI/GL report no consistent A1C effect and mixed fasting-glucose results, and that inconsistent definitions of high/low GI leave its clinical utility uncertain; emphasizes overall eating pattern and carbohydrate quality. 10.2337/dci19-0014
- Thomas, D., & Elliott, E. J. (2009). Low glycaemic index, or low glycaemic load, diets for diabetes mellitus. Cochrane Database of Systematic Reviews, (1), CD006296. Eleven RCTs (402 participants); low-GI diets lowered HbA1c by ~0.5 percentage points (WMD -0.5%, 95% CI -0.9 to -0.1) without increasing hypoglycaemia. 10.1002/14651858.CD006296.pub2
- Chekima, K., Yan, S. W., Lee, S. W. H., Wong, T. Z., Noor, M. I., Ooi, Y. B. H., et al. (2023). Low glycaemic index or low glycaemic load diets for people with overweight or obesity. Cochrane Database of Systematic Reviews, (6), CD005105. Updated review of 10 studies (1210 participants); concluded there is insufficient evidence to draw firm conclusions on weight, with certainty rated moderate to very low. Supersedes the more positive 2007 conclusion. 10.1002/14651858.CD005105.pub3
- Gardner, C. D., Trepanowski, J. F., Del Gobbo, L. C., Hauser, M. E., Rigdon, J., Ioannidis, J. P. A., et al. (2018). Effect of low-fat vs low-carbohydrate diet on 12-month weight loss in overweight adults and the association with genotype pattern or insulin secretion: the DIETFITS randomized clinical trial. JAMA, 319(7), 667-679. 609 adults, 12 months; no significant weight-loss difference between healthy low-fat and healthy low-carbohydrate diets, and neither genotype pattern nor baseline insulin secretion predicted which diet worked better. 10.1001/jama.2018.0245
- Reynolds, A., Mann, J., Cummings, J., Winter, N., Mete, E., & Te Morenga, L. (2019). Carbohydrate quality and human health: a series of systematic reviews and meta-analyses. Lancet, 393(10170), 434-445. Highest vs lowest dietary-fibre intake associated with 15-30% lower all-cause and cardiovascular mortality. 10.1016/S0140-6736(18)31809-9
- U.S. Department of Agriculture & U.S. Department of Health and Human Services. (2020). Dietary Guidelines for Americans, 2020-2025 (9th ed.). www.dietaryguidelines.gov/sites/default/files/2020-12/Dietary_Guidelines_for_Americans_2020-2025.pdf