Physical Address
304 North Cardinal St.
Dorchester Center, MA 02124
ISO 26642:2010 — Food Products — Determination of the Glycaemic Index (GI)
Standardised in vivo method for determining the glycaemic index of carbohydrate-containing foods
1. Standardised Protocol for GI Determination
ISO 26642:2010 establishes the internationally recognised method for determining the glycaemic index (GI) of carbohydrate-containing foods, developed by ISO/TC 34, Food products. The GI quantifies the blood glucose-raising ability of digestible carbohydrates by measuring the incremental area under the blood glucose response curve (IAUC) after consumption of a test food, expressed as a percentage of the response to an equivalent amount of reference carbohydrate (glucose, GI = 100). The standard is based on a Joint FAO/WHO Expert Consultation on Carbohydrates in Human Nutrition and represents the consensus of international experts in human nutrition and metabolism.
The standard requires a minimum of 10 healthy subjects, each tested on separate occasions with both the reference food (at least two and preferably three times) and the test food. Blood samples are taken at 0, 15, 30, 45, 60, 90, and 120 minutes, and the IAUC is calculated using the trapezoid rule. The final GI is expressed as the mean +/- standard error, rounded to the nearest whole number. The standard provides detailed sample calculations in Annex C, with worked examples showing IAUC computation for three reference replicates and one test food across 12 subjects. The trapezoid rule divides the postprandial period into segments, calculating each segment area as the average glucose rise multiplied by the time interval, then summing all positive increments only.
| GI Classification | Range | Examples |
|---|---|---|
| Low GI | <= 55 | Legumes, lentils, most vegetables, nuts |
| Medium GI | 56 – 70 | Whole wheat bread, basmati rice, sweet potato |
| High GI | > 70 | White bread, glucose, corn flakes, watermelon |
2. Subject Selection and Test Conditions
Rigorous subject management is central to reliable GI testing. Subjects must have no known history of diabetes, no medications affecting glucose tolerance (except stable contraceptives and certain chronic medications like thyroxin or antihypertensives), and no major medical events within the preceding 3 months. The fasting period is at least 10 hours, with no alcohol the previous evening and no vigorous exercise on test mornings. Capillary blood sampling is preferred over venous sampling because it yields less variable results and better discriminates between foods. Two fasting blood samples are taken and averaged as baseline. Subjects act as their own controls by consuming both reference and test foods on separate days, with the reference tested at least twice to establish individual glycaemic responsiveness.
The standard also specifies detailed criteria for subject exclusion and data handling. If a subject’s reference GI value falls outside the expected range or their fasting glucose is abnormal on a test day, that session must be repeated. Outliers are defined as values exceeding the mean +/- 2 standard deviations and may be excluded provided the minimum subject count remains at 10. The standard also addresses the handling of missing data points and the important consideration of subject drop-out during multi-visit protocols.
The mean within-subject coefficient of variation (CV) for the reference food must be <= 30%. Many small glucometer devices have analytical CVs above 3.6% and are explicitly deemed unsuitable for GI testing under this standard. The laboratory must use spectrophotometric or electrochemical enzyme systems with documented analytical performance, such as the YSI 2300 STAT Plus glucose analyser.
Capillary finger-prick sampling is the preferred method because it captures rapid postprandial glucose changes more accurately than venous or forearm site sampling, resulting in larger and more statistically detectable differences between foods. The test food must be consumed within 12-15 minutes at an even pace, and subjects should remain sedentary throughout the 2-hour test period.
3. Engineering Design Insights for GI Testing Laboratories
Setting up a compliant GI testing facility requires careful attention to several design factors. The laboratory must have a food preparatory area physically separated from the blood sampling area to prevent contamination and ensure subject comfort. Blood glucose analysis must use spectrophotometry or electrochemical enzyme systems with inter-assay CV < 3.6% on standard solutions. The instrument must be calibrated according to manufacturer specifications. Temperature control in the testing area is important as ambient temperature can affect peripheral blood flow and capillary glucose measurements.
The standard also addresses practical challenges such as testing multiple product flavours. If two flavours have essentially identical macronutrient composition, they may be tested in a single group of at least 5 subjects per flavour, with the combined result (>=10 subjects) reported as the GI for both. However, if statistically different GI values emerge (p < 0.05), individual flavours must be tested independently with 10 or more subjects each. The choice of reference food can significantly impact results: white bread has a GI of 71 relative to glucose, and conversion factors must be applied when using alternative reference foods.
Statistical analysis is a crucial component of GI determination. The standard recommends reporting the mean GI, standard error, and coefficient of variation. The 95% confidence interval for the GI value provides a measure of precision, and the standard recommends that products with GI values close to classification boundaries (55 and 70) be interpreted with appropriate caution given the inherent biological variability. For health claims and regulatory submissions, the standard emphasises that GI testing must be conducted by laboratories with demonstrated proficiency in the method, including documented training records and participation in inter-laboratory comparison programmes.
GI testing is only appropriate for foods containing >=10g of glycaemic carbohydrate per serving. Low-digestibility carbohydrates (resistant starch, sugar alcohols like erythritol and mannitol, polydextrose, inulin) must not be intentionally counted in the 50 g (or 25 g) carbohydrate portion used for testing. The distinction between total carbohydrate and glycaemic carbohydrate is frequently misunderstood in product development teams.
The IAUC calculation method and subject selection criteria specified in this standard ensure that GI values are reproducible across different testing facilities, provided the protocol is followed rigorously. The standard has become the global benchmark for glycaemic index determination in the food industry.
The IAUC calculation methodology and subject selection criteria defined in this standard have become the global benchmark for glycaemic index testing in the food industry. The trapezoid rule for IAUC calculation ensures consistent and reproducible results across different testing laboratories worldwide.
Understanding the distinction between total carbohydrate and glycaemic carbohydrate content is essential for proper application of this standard in product development and regulatory compliance contexts.
4. Frequently Asked Questions
Q1: Can GI values be calculated mathematically from ingredients?
No. The GI of heterogeneous foods can only be determined by in vivo testing per ISO 26642. Mathematical calculation from individual ingredients is explicitly prohibited because food matrix effects significantly influence glycaemic response, and synergistic interactions between components cannot be predicted.
No. The GI of heterogeneous foods can only be determined by in vivo testing per ISO 26642. Mathematical calculation from individual ingredients is explicitly prohibited because food matrix effects significantly influence glycaemic response, and synergistic interactions between components cannot be predicted.
Q2: When must a product be re-tested?
Re-testing is required when macronutrient composition changes (fat, protein, carbohydrate ratios), processing methods change, or physical/chemical factors (concentration, osmolality, acidity) are modified. Minor formulation changes and seasonal ingredient variations do not necessitate re-testing.
Re-testing is required when macronutrient composition changes (fat, protein, carbohydrate ratios), processing methods change, or physical/chemical factors (concentration, osmolality, acidity) are modified. Minor formulation changes and seasonal ingredient variations do not necessitate re-testing.
Q3: What is the minimum amount of glycaemic carbohydrate for GI testing?
The test food must contain 50 g of glycaemic carbohydrate (or 25 g if the 50 g portion is unreasonably large, e.g., for fruits). Foods with less than 10 g per serving should not be tested.
The test food must contain 50 g of glycaemic carbohydrate (or 25 g if the 50 g portion is unreasonably large, e.g., for fruits). Foods with less than 10 g per serving should not be tested.
Q4: Why must breakfast cereals be tested with water not milk?
Addition of milk influences the final GI of some products but not others. To ensure standardisation and cross-product comparability, the standard mandates using water for all cereal and powdered beverage tests. If milk is relevant to the product’s typical consumption, this can be noted separately.
Addition of milk influences the final GI of some products but not others. To ensure standardisation and cross-product comparability, the standard mandates using water for all cereal and powdered beverage tests. If milk is relevant to the product’s typical consumption, this can be noted separately.
