What Is the Glycemic Index?
The glycemic index (GI) is a numerical scale from 0 to 100 that ranks carbohydrate-containing foods according to the speed and magnitude of the blood glucose rise they produce after consumption, measured relative to a reference food (pure glucose or white bread, both set at 100). Foods that are digested and absorbed rapidly, causing a swift spike in blood glucose, receive high GI values. Foods that are digested more slowly, producing a gradual and sustained rise, receive low GI values.
GI classification per ISO 26642:2010 and the FAO/WHO framework:
|
GI category |
GI range |
Typical examples |
Glycemic Load (GL) per serving |
|
Low GI |
≤55 |
Lentils (~32), yogurt (~36), apple (~38), rolled oats (~55) |
Low GL: ≤10 — e.g., apple GL ≈ 6 |
|
Medium GI |
56–69 |
Basmati rice (~58), pita bread (~57), banana (~62), raisins (~64) |
Medium GL: 11–19 — varies by serving size |
|
High GI |
≥70 |
White bread (~75), instant oats (~79), white rice (~73), watermelon (~72) |
High GL: ≥20 — or low GL if serving is small (e.g., watermelon GL ≈ 4) |
GI values in this table are approximate and may vary by variety, preparation, ripeness, and processing. Authoritative reference values are published in the International Tables of Glycemic Index and Glycemic Load Values (Atkinson et al., 2021, American Journal of Clinical Nutrition), which lists over 4,000 foods tested under ISO-comparable methodology.
Why GI Testing Matters
- Diabetes management: Low-GI foods produce more gradual postprandial glucose excursions, supporting blood sugar management in type 1 and type 2 diabetes. Clinical dietary guidelines from Diabetes Canada, the American Diabetes Association, and European bodies include GI as a useful meal-planning tool.
- Weight management and satiety: Lower-GI foods tend to promote greater satiety for equivalent carbohydrate portions, potentially supporting appetite control and caloric intake regulation.
- Cardiovascular and metabolic disease risk: Epidemiological data link habitual low-GI dietary patterns with reduced risk of type 2 diabetes, coronary heart disease, and metabolic syndrome.
- Food product development: Manufacturers use GI testing to develop, validate, and market low-GI food products. Accurate GI values from ISO 26642:2010-compliant testing are required for any GI-based health or labeling claim in most international markets.
- Regulatory compliance: Several national regulatory frameworks (Australia, Canada, EU) have specific guidance on GI claims. ISO 26642:2010 is referenced as the required methodology in most.
Glycemic Index and Glycemic Load: Understanding Both Metrics
The glycemic index tells you how fast a carbohydrate raises blood glucose — but it doesn’t account for how much carbohydrate a typical serving contains. Glycemic Load (GL) integrates both dimensions, providing a more practical measure of a food’s actual impact on blood glucose at realistic portion sizes.
GL formula: GL = (GI × grams of available carbohydrate per serving) ÷ 100
- Low GL: ≤10 per serving
- Medium GL: 11–19 per serving
- High GL: ≥20 per serving
The practical importance of GL is illustrated by watermelon: its GI is approximately 72 (high), which might suggest it is a poor choice for glycemic management. But a typical 120 g serving of watermelon contains only about 6 g of available carbohydrate, giving a GL of approximately (72 × 6) ÷ 100 = 4.3, which is very low. Eating a normal portion of watermelon has a minimal real-world effect on blood glucose. In contrast, a large serving of white rice with a moderate GI can produce a high GL. Both GI and GL are routinely reported together in published GI tables and are tested in the same food science and nutritional analysis programs.
The International Standard: ISO 26642:2010
ISO 26642:2010 (Food products — Determination of the Glycaemic Index (GI) and Recommendation for Food Classification) is the definitive international methodology for GI measurement, developed in collaboration with the Food and Agriculture Organization (FAO) and the World Health Organization (WHO). It is the standard referenced by regulatory agencies, accredited laboratories, and the International Tables of Glycemic Index and Glycemic Load Values for published GI data.
ISO 26642:2010 establishes the following key requirements:
- Minimum 10 healthy subjects with normal fasting blood glucose and no metabolic disorders. A fasting plasma glucose <6.1 mmol/L (110 mg/dL) is typically required.
- Reference food tested at least twice per subject (preferably three times) on separate occasions under identical conditions. Multiple reference food tests are used to calculate within-subject variability.
- Within-subject CV ≤30% for reference food iAUC: the mean coefficient of variation across subjects for repeated reference food tests must not exceed 30%. If it does, one outlying reference test result per subject may be excluded, provided the subject has done the reference test at least three times.
- Carbohydrate dose: 50 g of available glycaemic carbohydrate (or 25 g for foods where a 50 g portion would be unreasonably large, such as some fruits). Non-digestible carbohydrates (resistant starch, some sugar alcohols, polydextrose) are not intentionally counted toward this 50 g.
- Minimum carbohydrate content: ≥10 g glycaemic carbohydrate per regular serving. Foods with less than 10 g available carbohydrate per serving should not be tested for GI.
- Blood glucose sampling schedule: fasting baseline measured at −10 and −5 minutes; postprandial samples at 15, 30, 45, 60, 90, and 120 minutes after food consumption begins.
- iAUC calculated by the trapezoid rule — the incremental area under the blood glucose response curve, computed geometrically and ignoring any area below the fasting baseline value.
- Food consumed within 10–15 minutes. Water (250 mL) is permitted with the test food; no other beverages during the test.
A critical regulatory point from ISO 26642: results from in vitro methods of assessing carbohydrate digestion should not be referred to as GI values. Only in vivo human testing per ISO 26642 can produce results that qualify as GI values for labeling and marketing claim purposes.
In Vivo GI Testing (ISO 26642:2010): Step-by-Step Protocol
Subject preparation and screening
Subjects fast for 10–12 hours before each test session, avoiding alcohol, caffeine, and strenuous exercise in the preceding period. Subjects with diabetes, metabolic syndrome, or other conditions affecting glucose metabolism are excluded. Fasting blood glucose is confirmed within normal range on each test day before the test food is administered.
Reference food testing
Each subject consumes the reference food — pure glucose (GI = 100 by definition) or white bread (see note on conversion below) — on at least two, preferably three, separate occasions under identical conditions. These repeated reference tests allow calculation of within-subject variability (CV). The reference food iAUC values form the denominator of the GI calculation.
If white bread is used as the reference food, because white bread has a GI of approximately 71 on the glucose scale, all GI values derived using white bread as reference (set at 100) must be multiplied by 0.71 to convert to glucose-equivalent GI values for standard international reporting and comparison.
Test food administration and blood sampling
Each subject consumes the test food portion providing 50 g of available carbohydrate within 10–15 minutes on a separate day. A capillary blood glucose sample (fingerstick) or venous blood sample is taken at the fasting baseline (averaged from −10 and −5 min readings) and then at 15, 30, 45, 60, 90, and 120 minutes postprandially. Portable glucose meters, continuous glucose monitors (CGMs), or laboratory hexokinase enzymatic analyzers may be used; laboratory analyzers provide the highest accuracy and are preferred for regulatory submissions.
iAUC calculation — the correct method
The incremental Area Under the Curve (iAUC) is calculated using the trapezoid rule applied to the blood glucose time-course, with the fasting baseline as the lower boundary. Critically, any portion of the curve that falls below the fasting baseline is excluded from the iAUC calculation — only the area above baseline is summed. This is the key distinction between iAUC (used by ISO 26642) and total AUC (which would include fasting glucose levels and gives different, non-comparable results between subjects with different fasting glucose levels).
GI calculation — the correct formula
GI = (iAUC for test food ÷ iAUC for reference food) × 100
Where: iAUC for test food = the incremental area under the subject’s blood glucose response curve for the test food; iAUC for reference food = the mean of the subject’s incremental areas from their two or three reference food test sessions.
Each subject’s individual GI is calculated first; the study GI is the mean of all subjects’ individual GIs. Individual GI values more than 2 standard deviations above the study mean may be treated as outliers and excluded, per ISO 26642 protocol.
Common error to note: Dividing reference iAUC by test food iAUC — inverting the numerator and denominator — produces a value that appears mathematically plausible but ranks foods in the opposite order from the correct GI scale. The test food is always the numerator.
In Vitro GI Estimation: Role, Methods, and Labeling Limitations
In vitro methods simulate carbohydrate digestion in a controlled laboratory setting, using digestive enzymes (salivary and pancreatic amylase, amyloglucosidase) to release glucose from the test food under standardized pH, temperature, and mechanical conditions. The rate and extent of glucose release from the food matrix over time are measured and compared to a glucose standard to generate an estimated GI value.
In vitro GI estimation is useful for:
- Rapid product development screening — comparing formulations before committing to expensive human trials
- Research on the mechanisms of GI modification (e.g., effects of fiber type, particle size, cooking method)
- Preliminary ingredient selection and process optimization
- Ranking foods within a category when relative GI differences are the objective
However, in vitro results cannot be used for GI labeling claims or GI-based marketing. ISO 26642:2010 explicitly states that in vitro results “should not be referred to as GI values”. This distinction is critical for food manufacturers: a product developed using in vitro GI estimation for formulation guidance still requires ISO 26642:2010-compliant human testing before any GI value can be placed on packaging or used in advertising claims. In vitro results are also frequently poor predictors of in vivo GI for complex food matrices — particularly those where food structure, viscosity, or matrix effects significantly influence gastric emptying and absorption rate.
See our companion article on in vivo and in vitro testing for a broader discussion of the distinction between human-based and laboratory-based testing methodologies.
Factors That Affect a Food’s Glycemic Index
A food’s GI value is not a fixed property of its chemical composition alone — it is profoundly influenced by processing, preparation, food matrix interactions, and ripeness. Understanding these factors is important both for interpreting GI data and for designing lower-GI food products.
- Starch structure (amylose vs amylopectin): Foods high in amylose (a long, linear chain) have lower GI than equivalent foods dominated by amylopectin (branched, rapidly digested). High-amylose rice and pasta have lower GI than standard varieties.
- Particle size and processing: Finely milled or heavily processed foods expose more starch surface area to digestive enzymes, increasing digestion rate and GI. Stone-ground whole wheat bread has a lower GI than finely milled white bread from the same flour, even with similar fiber content.
- Dietary fiber: Viscous, soluble fibers (beta-glucan in oats; pectin in fruits; psyllium) slow gastric emptying, reduce small intestinal absorption rate, and lower GI. Insoluble fiber has less effect on the GI.
- Cooking and food preparation: Gelatinization of starch during cooking increases digestibility and GI. Al dente pasta has a lower GI than fully cooked pasta from the same flour. Cooling cooked starch promotes retrogradation (formation of resistant starch), which lowers GI — explaining why cold cooked potatoes have lower GI than freshly hot cooked potatoes.
- Protein and fat content: The presence of protein and fat in a meal slows gastric emptying and reduces the glycemic response to co-consumed carbohydrates. However, GI testing measures the food eaten alone under standardized conditions — mixed-meal GI is more complex.
- Acidity: Acidic foods and ingredients (vinegar, lemon juice, sourdough fermentation, citric acid) slow gastric emptying and starch hydrolysis, reducing the glycemic response. Sourdough bread has a substantially lower GI than equivalent non-fermented bread.
- Ripeness: As fruits ripen, starches convert to sugars and GI rises. Unripe bananas have a GI ~30–40; fully ripe bananas reach ~60–70.
- Food variety and origin: Different varieties of nominally the same food can have markedly different GI values — Basmati rice (~58) vs. jasmine rice (~109); waxy vs. standard potato varieties.
These factors mean that GI testing should be conducted on the food in the form it will be consumed, as specified by ISO 26642:2010. A GI value obtained from a tested bread sample applies to that specific bread baked under those specific conditions — reformulation, ingredient changes, or process modifications may require retesting.
Finding Accredited GI Testing Laboratories
ISO 26642:2010-compliant GI testing requires a specialized research facility capable of: recruiting and screening healthy human subjects under IRB/ethics board oversight; standardizing and preparing test foods; conducting multiple-session crossover protocols; collecting and analyzing capillary or venous blood samples under controlled conditions; and calculating iAUC values by the trapezoid rule with outlier handling. These are clinical research capabilities, not standard food chemistry lab capabilities.
A full ISO 26642:2010 GI study for a single food typically requires 10–20 subjects, 3–5 visits per subject (reference food tested 2–3 times plus test food), and several weeks of subject recruitment and data collection, with a corresponding laboratory cost and timeline. Turnaround time for a complete ISO 26642 GI study at accredited contract labs typically ranges from 8 to 16 weeks. In vitro GI screening can be arranged through food chemistry and nutritional analysis laboratories on shorter timelines and at lower cost.
Contract Laboratory connects food and beverage manufacturers and nutritional researchers with accredited food science and nutritional analysis laboratories and clinical research facilities experienced in GI testing programs. See also our companion GI resource for additional background on glycemic index testing in food science.
Conclusion
Glycemic index testing is a specialized intersection of clinical research methodology and food science — requiring in vivo human protocols under ISO 26642:2010 to produce the GI values that can be used for labeling, health claims, and published databases. Understanding the correct formula (iAUC test food ÷ iAUC reference food × 100), the distinction between iAUC and total AUC, the role of reference food choice, and the limitations of in vitro methods are all essential for anyone commissioning, interpreting, or publishing GI testing data. Glycemic Load extends GI by incorporating serving-size carbohydrate content, providing a more complete picture of real-world glycemic impact.
Contract Laboratory connects food manufacturers, nutritional researchers, and supplement brands with accredited food science and nutritional analysis laboratories and clinical research facilities capable of conducting ISO 26642:2010-compliant GI testing programs.
Submit a testing request or contact our team.
This article was created with the assistance of Generative AI and has undergone editorial review before publishing.
Frequently Asked Questions About Glycemic Index Testing
GI = (iAUC for the test food ÷ iAUC for the reference food) × 100. The test food’s incremental blood glucose area under the curve (iAUC) is the numerator, divided by the mean iAUC of the reference food (tested on separate occasions by the same subjects), multiplied by 100. A food with a smaller blood glucose response than glucose has a GI below 100; a food with a larger response would exceed 100. The iAUC is the incremental area — calculated by the trapezoid rule above the fasting baseline, excluding any area where the blood glucose dips below baseline.
ISO 26642:2010 is the international standard for GI measurement methodology, developed in collaboration with FAO and WHO. It specifies the complete in vivo human testing protocol: minimum 10 healthy subjects, reference food tested at least twice per subject, within-subject CV ≤30% for reference food iAUC, 50 g available carbohydrate dose, standardized blood sampling schedule (15, 30, 45, 60, 90, 120 minutes), and iAUC calculation by the trapezoid rule. ISO 26642:2010 matters because it is the methodology that regulatory bodies and the International Tables of GI Values recognize as producing valid, comparable GI values. In vitro methods, regardless of how well-designed, cannot produce results described as GI values under ISO 26642.
Total AUC (area under the curve) includes the area under the entire blood glucose time-course from baseline to 120 minutes — including the area attributable to the fasting glucose concentration. Incremental AUC (iAUC), as specified by ISO 26642:2010, calculates only the area above the fasting baseline, ignoring any portions of the curve that fall below baseline. This distinction matters because subjects with different fasting glucose levels would give different total AUC values even if their postprandial glucose responses were identical, making cross-subject comparison unreliable. The iAUC normalizes for baseline differences and focuses solely on the glycemic response to the food, which is the biologically relevant quantity for GI calculation.
No. ISO 26642:2010 explicitly states that results from in vitro methods of assessing carbohydrate digestion should not be referred to as GI values. In vitro testing is useful for product development screening, ingredient comparison, and formulation research — but it cannot substitute for ISO 26642:2010-compliant human trials when GI values are to be stated on product labels or used in marketing claims. Most international regulatory frameworks that address GI labeling reference ISO 26642 or equivalent in vivo methodology as the required standard of evidence. Manufacturers who have developed a product using in vitro GI estimation for formulation guidance still require a formal ISO 26642 human study before making a GI claim.
Glycemic index measures how rapidly a carbohydrate raises blood glucose relative to a reference food — it is a property of the food’s carbohydrate quality. Glycemic load (GL) accounts for both carbohydrate quality and quantity: GL = (GI × grams of available carbohydrate per serving) ÷ 100. A low GL is ≤10; medium is 11–19; high is ≥20. GL is often more practically useful because it captures the real-world impact of a normal portion. Watermelon has a high GI (~72) but a low GL (~4) per standard serving because a typical portion is mostly water with very few grams of carbohydrate. Both metrics are routinely reported together in GI testing programs and food labeling guidance.
ISO 26642:2010 requires a minimum of 10 healthy human subjects. All subjects must have normal fasting blood glucose and no metabolic disorders affecting glucose response. The reference food must be tested at least twice per subject (preferably three times) on separate days under identical conditions. The mean within-subject coefficient of variation for reference food iAUC across the group must not exceed 30%; if it does, one outlying reference result per subject may be excluded if the subject completed at least three reference tests. In practice, many commercial GI testing laboratories recruit 10–15 subjects to allow for dropout and data exclusions while maintaining sufficient power.
Pure glucose is set at GI = 100 by definition and is the preferred reference food under ISO 26642:2010. White bread is sometimes used instead, particularly for starchy foods where comparison to bread is more nutritionally meaningful. However, white bread itself has a GI of approximately 71 on the glucose scale. If white bread is used as the reference food (assigned a value of 100), all GI values produced in that study are expressed on the white-bread scale and must be multiplied by 0.71 to convert to the standard glucose-scale GI values for comparison with published tables. Failing to apply this conversion leads to inflated GI values that are not comparable to glucose-referenced data from other studies.
Starch structure is the primary determinant — foods high in amylose (linear starch) digest more slowly and have lower GI than amylopectin-dominant foods. Particle size matters significantly: finely milled flours digest faster than coarsely ground equivalents. Cooking increases GI by gelatinizing starch, while cooling promotes retrogradation (resistant starch formation), which lowers GI. Viscous soluble fiber (beta-glucan, pectin, psyllium) slows gastric emptying and reduces GI. Acidity — from vinegar, fermentation (sourdough), or naturally occurring organic acids — slows starch digestion and lowers GI. Fruit ripeness raises GI as starches convert to simple sugars. Because GI is so sensitive to these variables, ISO 26642:2010 requires testing the food in the exact form it will be consumed, and reformulation or process changes may require retesting.