Glycemic Index and Glycemic Load

The glycemic index is a number that ranks a carbohydrate food by how far it raises blood glucose in the two hours after eating, compared with the same amount of carbohydrate taken as pure glucose. Glycemic load combines that ranking with the amount of carbohydrate in a serving. Both are measures, not remedies. Pooled randomized trials in people with diabetes found that eating patterns built on low-glycemic-index foods lowered hemoglobin A1C (HbA1c), the blood test that reflects average glucose over about three months, by a small amount on top of usual treatment. Long-term observational studies link diets with a high index or load to a higher rate of new type 2 diabetes. The same body of research shows the weak point of the measure: the number for one food varies widely from person to person and from one test day to the next.

Evidence base: three pooled analyses of randomized trials in diabetes, two umbrella reviews, two pooled analyses in gestational diabetes, seven pooled analyses and reviews of prospective cohort studies, the original 1981 and 1983 test-meal studies, three editions of the international tables, two regional food-testing studies and three studies of person-to-person variation. Every figure on this page is taken from the published abstract of the study named beside it.


Table of Contents

  1. What the Glycemic Index Measures
  2. How It Is Measured
  3. What Glycemic Load Adds
  4. What Trials in Diabetes Found
  5. What Long-Term Studies of Diabetes Risk Found
  6. Why the Same Food Gives Different Readings
  7. What Studies of Personal Glucose Responses Found
  8. Reviews Written to Inform Guidelines
  9. Safety Findings
  10. Limits of the Evidence
  11. Key Research Papers
  12. Connections
  13. Featured Videos

What the Glycemic Index Measures

Starches and sugars are digested into glucose, which enters the blood. Some foods release their glucose quickly and some slowly. The glycemic index puts a number on that difference.

The idea comes from a 1981 feeding study in healthy volunteers (PMID 6259925). The researchers fed 62 commonly eaten foods and sugars, one at a time, to groups of 5 to 10 healthy fasting volunteers and measured blood glucose over 2 hours. Each result was expressed as a percentage of the response seen when the same amount of carbohydrate was taken as glucose. Glucose is therefore the yardstick, set at 100.

Average blood glucose response by food group in the 1981 study, as a percentage of the response to glucose. The groups are named as the abstract names them.
Food groupAverage response
Vegetables70 ± 5%
Breakfast cereals65 ± 5%
Cereals and biscuits60 ± 3%
Fruit50 ± 5%
Dairy products35 ± 1%
Dried legumes (beans, peas, lentils)31 ± 3%

In 1983 the same group ran a comparative test-meal study in people with diabetes (PMID 6688055). Groups of five to seven patients ate 50 g carbohydrate portions of 15 foods as breakfast test meals. The responses differed two- to threefold between foods. Spaghetti, a bran breakfast cereal, rice and beans gave responses significantly below bread, and a corn flake cereal gave a response above it. Across five legumes, the average peak rise in blood glucose and the average area under the glucose curve were 23% and 28% lower than the average for seven other starchy foods (p < 0.001). The authors wrote that these differences could not be predicted from tables of chemical composition.

That last point is why the measure exists. The authors of the 2002 international table described the glycemic index as a more useful concept than sorting carbohydrates by chemistry into “simple or complex” or “sugars or starches” (PMID 12081815).

The Low, Medium and High Categories

On the 100-point glucose scale, two food-testing studies state the categories they used: low is 55 or less and high is 70 or more (PMID 28532533), with medium given as 56 to 69 (PMID 40586370).

The 2021 edition of the international tables, a systematic review, found that dairy products, legumes, pasta and fruits were usually low-glycemic-index foods (55 or less) with consistent values around the world. Cereals and cereal products, including whole-grain versions, showed wide variation. Breads, breakfast cereals, rice and regional foods were available in high, medium and low versions. Most varieties of potato were high, although specific low-index varieties have been identified (PMID 34258626).

Back to Table of Contents

How It Is Measured

How a glycemic index value is calculated from two blood glucose curves One person, two test days, the same amount of carbohydrate 0 15 30 45 60 90 120 Minutes after eating (the horizontal line is the fasting glucose level) Blood glucose Glucose drink (the reference, set at 100) Test food Glycemic index = area under the test food curve ÷ area under the glucose curve × 100 Schematic curves to show the method. They are not measured data.

A 2017 food-testing study describes the procedure step by step (PMID 28532533):

  1. At least fifteen healthy participants tested each food, after an overnight fast.
  2. Each portion held 25 or 50 g of available carbohydrate, meaning the carbohydrate the body can digest and absorb.
  3. The reference food, glucose, was tested three times and the test food once, on separate occasions.
  4. Finger-prick blood samples were taken before eating and at 15, 30, 45, 60, 90 and 120 minutes afterwards.
  5. The value was calculated as the incremental area under the blood glucose curve (the area above the fasting level) for the test food, as a percentage of the same participant’s average area for the reference food.

Laboratories have not all worked the same way. A 2026 study of 50 cereal-based foods tested each product on at least 10 healthy participants and used both glucose and white bread as reference foods (PMID 40586370). The 2021 international tables keep two lists: about 2,100 items measured with the method recommended by the International Standards Organization, and about 1,900 items measured with less robust methods, including small numbers of healthy subjects (PMID 34258626).

The International Tables

Back to Table of Contents

What Glycemic Load Adds

The glycemic index is measured on a fixed amount of carbohydrate, not on the portion a person normally eats. A food can rank high and still deliver little glucose if a serving holds little carbohydrate. Glycemic load was introduced to cover that gap.

The 2002 international table defines glycemic load as the glycemic index multiplied by the carbohydrate content, and that edition also lists the load for specified serving sizes (PMID 12081815). The 2026 food-testing study calculated it as the product of a food’s index and the amount of available carbohydrate in a reference serving (PMID 40586370).

Because the index is a percentage, the load comes out in grams. As an illustration only, not a measured food: an index of 50 and a serving holding 20 g of available carbohydrate give a load of 10.

Loads can be added up over a whole day. Across 24 prospective cohort studies, the daily glycemic load ranged from about 60 to about 280 g per 2,000 kcal eaten (PMID 23364021).

The records read for this page state category cut-offs for the index but none for the load, so no low, medium or high bands for glycemic load are given here.

Back to Table of Contents

What Trials in Diabetes Found

In these trials people were assigned at random to an eating pattern built on low-glycemic-index foods or to a comparison diet. The diet was tested in addition to the participants’ usual treatment. None of the abstracts describes a test of diet as a replacement for medication.

The 2021 Meta-Analysis

A systematic review and meta-analysis of randomized controlled trials pooled 29 trial comparisons in 1,617 participants with type 1 and type 2 diabetes (PMID 34348965). The trials lasted three weeks or more. Participants were predominantly middle-aged, overweight or obese, with moderately controlled type 2 diabetes treated with glucose-lowering drugs or insulin.

The authors described the result as “small important improvements” obtained “beyond concurrent treatment with hyperglycaemia drugs or insulin”.

Other Pooled Analyses of Randomized Trials

A 2019 systematic review and meta-analysis pooled 54 randomized controlled trials in adults or children with impaired glucose tolerance, type 1 diabetes or type 2 diabetes (PMID 31374573). Low-glycemic-index diets reduced HbA1c, fasting glucose, body mass index, total cholesterol and low-density lipoprotein cholesterol. They had no effect on fasting insulin, on a calculated score of insulin resistance, on high-density lipoprotein cholesterol, on triglycerides or on insulin requirements. The abstract gives no pooled figure in HbA1c units. The reduction in fasting glucose and HbA1c was inversely correlated with body weight, and the greatest reduction in fasting glucose was seen in the longest studies. On triglycerides this review and the 2021 one disagree.

A 2023 network meta-analysis, a method that compares many diets at once through direct and indirect trial evidence, covered 42 randomized trials of at least 6 months in 4,809 people with type 2 diabetes (PMID 37513574). The low-glycemic-index diet was one of five diets that significantly lowered fasting glucose against a control diet; the abstract gives the figure as −1.26 (−2.26 to −0.27) without naming the unit. It was not among the three diets the abstract lists as significantly lowering HbA1c, although a combined ranking for both outcomes placed it among four diets with “promising effects”. The average falls in HbA1c and fasting glucose were significantly related only to how much weight participants lost.

A 2023 umbrella review (a review of published meta-analyses) of diet in type 2 diabetes, limited to trials of at least 12 weeks, found moderate certainty of evidence for beneficial effects of low-glycemic-index diets, among several other diets, on various cardiometabolic measures. It judged the methodological quality of 77% of the included publications as low to very low (PMID 38027413).

For body weight the finding is close to null. A 2022 umbrella review of weight-loss diets in adults with type 2 diabetes found that low-glycemic-index diets, like high-protein, Mediterranean and vegetarian diets, achieved minimal (0.3 to 2 kg) or no difference from control diets, in meta-analyses it rated as low to critically low in quality (PMID 34796367).

Gestational Diabetes

In gestational diabetes, a 2020 meta-analysis of six randomized trials in 532 women found that a low-glycemic-index diet lowered glucose measured 2 hours after meals (standardized mean difference −0.46, 95% confidence interval −0.82 to −0.10, p = 0.01). It found no substantial influence on fasting glucose, HbA1c, birth weight, the rate of large babies (macrosomia) or the need for insulin (PMID 30231782).

A 2024 network meta-analysis of 39 trials in 2,712 women also found an advantage for low-glycemic-index diets in 2-hour post-meal glucose (mean difference −6.3, 95% credible interval −9.9 to −2.7; the abstract names no unit) and, unlike the 2020 analysis, a lower birth weight (mean difference −180.09, 95% credible interval −267.48 to −94.65). It rated the certainty of both findings as low (PMID 39481539).

Back to Table of Contents

What Long-Term Studies of Diabetes Risk Found

Whether a high-index diet leads to type 2 diabetes cannot be tested in a short trial. The evidence comes from prospective cohort studies: large groups of people without diabetes report what they eat and are followed for years. Such studies show an association. They cannot prove that the diet caused the difference.

Results are given as a relative risk. A relative risk of 1.27 means the group with the highest-index diet developed type 2 diabetes at 1.27 times the rate of the group with the lowest. A relative risk of 1.00 means no difference.

Meta-analyses of prospective cohort studies: relative risk of type 2 diabetes, with the 95% confidence interval in brackets.
YearStudies pooledGlycemic indexGlycemic loadSource
200837 cohort studies covering several chronic diseases1.40 (1.23 to 1.59), highest against lowest intake1.27 (1.12 to 1.45)PMID 18326601
201321 cohort studies1.08 (1.02 to 1.15) per 5 index units1.03 (1.00 to 1.05) per 20 load unitsPMID 24265366
201324 cohort studiesNot reported1.45 (1.31 to 1.61) per 100 g of loadPMID 23364021
2014Three large United States cohorts plus other studies1.19 (1.14 to 1.24), highest against lowest category1.13 (1.08 to 1.17)PMID 24787496
2024Ten cohorts of 100,000 or more participants each1.27 (1.21 to 1.34), highest against lowest intake1.15 (1.09 to 1.21)PMID 38272606

The 2014 report followed 74,248 and 90,411 women in two cohorts of nurses and 40,498 men in a cohort of health professionals, and recorded 15,027 new cases of type 2 diabetes over 3,800,618 person-years. People in the highest fifth for glycemic index had a 33% higher risk than those in the lowest fifth (95% confidence interval 26% to 41%); for glycemic load the figure was 10% (2% to 18%). A diet high in index or load and low in cereal fibre carried about a 50% higher risk (PMID 24787496).

The 2024 meta-analysis of very large cohorts also found that diets high in fibre and whole grains showed associations similar to those of low-glycemic-index diets (PMID 38272606).

Results That Weaken the Picture

Back to Table of Contents

Why the Same Food Gives Different Readings

Variation Between People and Between Test Days

A 2016 clinical trial set out to measure how reliable a glycemic index value is (PMID 27604773). Sixty-three volunteers free from chronic disease, aged 18 to 85, each went through 3 sets of tests with glucose and white bread, both providing 50 g of available carbohydrate, using standardized methods.

The authors concluded that there is substantial variability in individual responses and that the glycemic index “is unlikely to be a good approach to guiding food choices”.

Variation Between Products With the Same Name

The international tables report that foods sharing one name can sit in different categories. The 2008 edition found breads, breakfast cereals and rice, including whole-grain kinds, in both high and low versions (PMID 18835944). The 2021 edition attributed the wide variation in cereal products, “presumably”, to differences in manufacturing methods (PMID 34258626).

Fat, Protein and Fibre

In the 1981 study, across the 62 foods, a higher fat content (p < 0.01) and a higher protein content (p < 0.001) went with a smaller rise in blood glucose. No such relationship was found with fibre or sugar content (PMID 6259925).

The 1983 study in people with diabetes tested three changes that were predicted to alter the response, and found all three “without effect”: replacing white bread with wholemeal bread, substantially raising the simple sugars in the meal, and doubling the protein in the meal by adding cottage cheese to bread (PMID 6688055). The two abstracts therefore do not agree on protein, and neither found an effect of fibre.

Ripeness, Cooking and Cooling

Ripeness, cooking time and cooling after cooking are often named as things that change a food’s value. None of the records read for this page measured them, so this page reports no figures for them. Starch that resists digestion after cooking and cooling is the subject of the Resistant Starch page.

Back to Table of Contents

What Studies of Personal Glucose Responses Found

Two large studies measured how differently people respond to the same meal.

A 2015 study continuously monitored glucose for a week in an 800-person cohort and measured responses to 46,898 meals (PMID 26590418). It found high variability in the response to identical meals. The researchers built a computer model from blood tests, dietary habits, body measurements, physical activity and gut bacteria that predicted each person’s response to real-life meals, and confirmed it in a separate 100-person cohort. A blinded randomized controlled dietary intervention based on the model then produced significantly lower post-meal glucose responses. The abstract gives no size for that reduction and no HbA1c result.

A 2020 study of 1,002 twins and unrelated healthy adults in the United Kingdom found that after identical meals the between-person variation in blood glucose response was 68%, expressed as a coefficient of variation (PMID 32528151). For blood glucose, the make-up of the meal still explained more of the variation (15.4%) than person-specific factors such as gut bacteria (6.0%). The findings were confirmed in a cohort of 100 people in the United States.

Both studies measured the person rather than the food. A table value is an average across the volunteers in a test; these studies show how far one person’s response can sit from that average. Measuring one’s own response is covered on the Glucose Monitoring page.

Back to Table of Contents

Reviews Written to Inform Guidelines

The sources read for this page do not include the text of any diabetes guideline, so no guideline statement is quoted here. Two of the reviews were written to feed into guidance, and they reached different judgements because they asked different questions.

The authors of the 2024 cohort meta-analysis wrote that dietary recommendations to reduce glycemic index and load could have effects on health outcomes similar to those of recommendations to increase fibre and whole grain intake (PMID 38272606).

Back to Table of Contents

Safety Findings

Back to Table of Contents

Limits of the Evidence

Back to Table of Contents

Key Research Papers

  1. Jenkins DJ, Wolever TM, Taylor RH, Barker H, Fielden H, Baldwin JM, et al. Glycemic index of foods: a physiological basis for carbohydrate exchange. Am J Clin Nutr. 1981;34(3):362-6. PubMed PMID: 6259925
  2. Jenkins DJ, Wolever TM, Jenkins AL, Thorne MJ, Lee R, Kalmusky J, et al. The glycaemic index of foods tested in diabetic patients: a new basis for carbohydrate exchange favouring the use of legumes. Diabetologia. 1983;24(4):257-64. PubMed PMID: 6688055
  3. Foster-Powell K, Holt SH, Brand-Miller JC. International table of glycemic index and glycemic load values: 2002. Am J Clin Nutr. 2002;76(1):5-56. PubMed PMID: 12081815
  4. Atkinson FS, Foster-Powell K, Brand-Miller JC. International tables of glycemic index and glycemic load values: 2008. Diabetes Care. 2008;31(12):2281-3. PubMed PMID: 18835944
  5. Atkinson FS, Brand-Miller JC, Foster-Powell K, Buyken AE, Goletzke J. International tables of glycemic index and glycemic load values 2021: a systematic review. Am J Clin Nutr. 2021;114(5):1625-1632. PubMed PMID: 34258626
  6. Al Dhaheri AS, Henry CJK, Mohamad MN, Ohuma EO, Ismail LC, Al Meqbaali FT, et al. Glycaemic index and glycaemic load values of commonly consumed foods in the United Arab Emirates. Br J Nutr. 2017;117(8):1110-1117. PubMed PMID: 28532533
  7. Khaliq M, Iqbal S, Saeed F, Sanghvi G, Ahmed F, Amjad A, et al. Determination of Glycemic Index and Glycemic Load of Indigenous Cereal Products of Pakistan. J Am Nutr Assoc. 2026;45(1):6-13. PubMed PMID: 40586370
  8. Chiavaroli L, Lee D, Ahmed A, Cheung A, Khan TA, Blanco S, et al. Effect of low glycaemic index or load dietary patterns on glycaemic control and cardiometabolic risk factors in diabetes: systematic review and meta-analysis of randomised controlled trials. BMJ. 2021;374:n1651. PubMed PMID: 34348965
  9. Zafar MI, Mills KE, Zheng J, Regmi A, Hu SQ, Gou L, et al. Low-glycemic index diets as an intervention for diabetes: a systematic review and meta-analysis. Am J Clin Nutr. 2019;110(4):891-902. PubMed PMID: 31374573
  10. Jing T, Zhang S, Bai M, Chen Z, Gao S, Li S, et al. Effect of Dietary Approaches on Glycemic Control in Patients with Type 2 Diabetes: A Systematic Review with Network Meta-Analysis of Randomized Trials. Nutrients. 2023;15(14):3156. PubMed PMID: 37513574
  11. Szczerba E, Barbaresko J, Schiemann T, Stahl-Pehe A, Schwingshackl L, Schlesinger S. Diet in the management of type 2 diabetes: umbrella review of systematic reviews with meta-analyses of randomised controlled trials. BMJ Med. 2023;2(1):e000664. PubMed PMID: 38027413
  12. Churuangsuk C, Hall J, Reynolds A, Griffin SJ, Combet E, Lean MEJ. Diets for weight management in adults with type 2 diabetes: an umbrella review of published meta-analyses and systematic review of trials of diets for diabetes remission. Diabetologia. 2022;65(1):14-36. PubMed PMID: 34796367
  13. Xu J, Ye S. Influence of low-glycemic index diet for gestational diabetes: a meta-analysis of randomized controlled trials. J Matern Fetal Neonatal Med. 2020;33(4):687-692. PubMed PMID: 30231782
  14. Zhang L, Wang F, Tashiro S, Liu PJ. Effects of Dietary Approaches and Exercise Interventions on Gestational Diabetes Mellitus: A Systematic Review and Bayesian Network Meta-analysis. Adv Nutr. 2024;15(12):100330. PubMed PMID: 39481539
  15. Barclay AW, Petocz P, McMillan-Price J, Flood VM, Prvan T, Mitchell P, et al. Glycemic index, glycemic load, and chronic disease risk--a meta-analysis of observational studies. Am J Clin Nutr. 2008;87(3):627-37. PubMed PMID: 18326601
  16. Greenwood DC, Threapleton DE, Evans CE, Cleghorn CL, Nykjaer C, Woodhead C, et al. Glycemic index, glycemic load, carbohydrates, and type 2 diabetes: systematic review and dose-response meta-analysis of prospective studies. Diabetes Care. 2013;36(12):4166-71. PubMed PMID: 24265366
  17. Livesey G, Taylor R, Livesey H, Liu S. Is there a dose-response relation of dietary glycemic load to risk of type 2 diabetes? Meta-analysis of prospective cohort studies. Am J Clin Nutr. 2013;97(3):584-96. PubMed PMID: 23364021
  18. Bhupathiraju SN, Tobias DK, Malik VS, Pan A, Hruby A, Manson JE, et al. Glycemic index, glycemic load, and risk of type 2 diabetes: results from 3 large US cohorts and an updated meta-analysis. Am J Clin Nutr. 2014;100(1):218-32. PubMed PMID: 24787496
  19. Jenkins DJA, Willett WC, Yusuf S, Hu FB, Glenn AJ, Liu S, et al. Association of glycaemic index and glycaemic load with type 2 diabetes, cardiovascular disease, cancer, and all-cause mortality: a meta-analysis of mega cohorts of more than 100 000 participants. Lancet Diabetes Endocrinol. 2024;12(2):107-118. PubMed PMID: 38272606
  20. Reynolds A, Mann J, Cummings J, Winter N, Mete E, Te Morenga L. Carbohydrate quality and human health: a series of systematic reviews and meta-analyses. Lancet. 2019;393(10170):434-445. PubMed PMID: 30638909
  21. Jayedi A, Soltani S, Jenkins D, Sievenpiper J, Shab-Bidar S. Dietary glycemic index, glycemic load, and chronic disease: an umbrella review of meta-analyses of prospective cohort studies. Crit Rev Food Sci Nutr. 2022;62(9):2460-2469. PubMed PMID: 33261511
  22. Matthan NR, Ausman LM, Meng H, Tighiouart H, Lichtenstein AH. Estimating the reliability of glycemic index values and potential sources of methodological and biological variability. Am J Clin Nutr. 2016;104(4):1004-1013. PubMed PMID: 27604773
  23. Zeevi D, Korem T, Zmora N, Israeli D, Rothschild D, Weinberger A, et al. Personalized Nutrition by Prediction of Glycemic Responses. Cell. 2015;163(5):1079-1094. PubMed PMID: 26590418
  24. Berry SE, Valdes AM, Drew DA, Asnicar F, Mazidi M, Wolf J, et al. Human postprandial responses to food and potential for precision nutrition. Nat Med. 2020;26(6):964-973. PubMed PMID: 32528151

PubMed Topic Searches

  1. PubMed: Low-glycemic-index diets in diabetes, meta-analyses
  2. PubMed: Glycemic load and risk of type 2 diabetes, cohort studies
  3. PubMed: International tables of glycemic index and glycemic load
  4. PubMed: Reliability and variability of glycemic index values
  5. PubMed: Personal glucose responses to food

Back to Table of Contents

Connections

Back to Table of Contents