A glycaemic index (GI) chart like the two below gives one number per food, which can make the list look more exact than it is. Before using one to plan breakfast or a snack, it helps to know how much that number can move. This page puts two charts side by side. The first shows the same six foods measured in three laboratories on three continents. The second shows seventeen date varieties measured in one laboratory in Saudi Arabia.
How GI is measured
A volunteer arrives after an overnight fast and eats a portion of the test food that holds 50 g of available carbohydrate. Finger-prick blood samples follow over two hours. The rise in blood glucose is then compared with the rise after 50 g of glucose, which is set at 100. The three-lab paper below uses the standard bands: 55 or less is low GI, 56 to 69 is medium, and 70 or more is high.
Chart 1: six foods, three labs
Wolever and colleagues sent the same batches of six commercial cereal products to the University of Sydney in Australia, GI Labs in Toronto in Canada, and Biofortis Mérieux NutriSciences in Saint-Herblain in France. All three labs followed the 2010 ISO method and one shared protocol, testing healthy non-smokers aged 18 to 35 with a BMI between 19 and 25. The paper numbers the labs 1 to 3 without saying which is which. Values are from its Table 3.
| Food | Lab 1 | Lab 2 | Lab 3 | All labs (mean ± SD) |
|---|---|---|---|---|
| Rotary-moulded biscuit | 48 | 47 | 45 | 47 ± 18 |
| Sandwiched rotary-moulded biscuit | 44 | 56 | 52 | 50 ± 19 |
| Cracker | 51 | 61 | 60 | 57 ± 18 |
| White bread | 76 | 65 | 63 | 68 ± 22 |
| Corn flakes | 77 | 77 | 66 | 74 ± 21 |
| Gingerbread | 81 | 88 | 89 | 86 ± 23 |
Look along each row. White bread came out at 76 in one lab and 63 in another, which is high GI in the first and medium in the second. Corn flakes did the same. The cracker and the sandwiched biscuit crossed the line between low and medium. Gingerbread was high in all three labs and the plain biscuit was low in all three.
The authors still read this as good news for the method. The lab averages did not differ significantly (p = 0.57), and no food was rated high by one lab and low by another. They put the typical lab-to-lab spread at a standard deviation of 5.1. From that, they calculate that when exactly the same food is tested twice by the ISO method, the 95% confidence interval for the difference between the two results is about ±14 points. The study was funded by Mondelez International, two of the authors work there, and the paper says the funders helped design the study and write it up but had no role in collecting or analysing the data.
Chart 2: seventeen dates, one lab
AlGeffari and colleagues at the College of Medicine, Qassim University, tested seventeen tamar-stage varieties on healthy adults, ten men and nine women. Sixteen of the varieties were grown in Al-Qassim, and the Ajwah sample came from Al-Madinah. Each date was eaten once, by nine or ten volunteers. The rows below are a selection from the paper's Table 2: the two highest, the two lowest, plus Ajwah and Medjool.
| Variety | Grams eaten | GI (mean ± SEM) | GL |
|---|---|---|---|
| Sellaj | 68.96 | 74.6 ± 10.1 | 24 |
| Maktoomi | 68.51 | 71.0 ± 11.1 | 16.1 |
| Ajwah | 70.13 | 55.9 ± 5.9 | 8.5 |
| Medjool | 70.54 | 55.3 ± 6.8 | 17.2 |
| Sukkary | 77.63 | 43.4 ± 4.7 | 11.7 |
| Shaqra | 66.92 | 42.8 ± 5.5 | 9.2 |
Across all seventeen the GI ran from 42.8 to 74.6, a gap of almost 32 points, yet the authors found no significant difference in GI between varieties. The standard errors are wide, up to 12.1 for Osilah. GL told a different story: it ranged from 8.5 for Ajwah to 24 for Sellaj, and those differences were significant (P < .001). Ajwah at 55.9 and Medjool at 55.3 sit right on the line between low and medium, so treat both as borderline.
What the two charts say together
Chart 1 shows that with the same batch and the same method, a GI can still shift by more than ten points between labs. Chart 2 shows dates spread over a wide range, but the study could not show that any variety's GI differed from the others. Neither paper tested Sayer, so no number here belongs to it. Sayer is a semi-dry variety from Khuzestan in Iran, and it is also traded as Stamaran.
This page reports published test results and is not dietary or medical advice. Sources: Wolever TMS et al., Nutrients 2019 (PMID 31540317); AlGeffari MA et al., Ann Saudi Med 2016 (PMID 27920411). More articles are on the blog.
More about this site is on the About Us page and the contact page.