The Glycaemic Index Of Foods: A Physiological Basis For Carbohydrate Exchange
It is with gratitude and thanks that I send out this long-overdue newsletter to update you on some issues in the nutritional world. I also humbly apologise for my absence with MISSION: NUTRITION, but resources and time constraints due to my post-graduate studies have been a hurdle. My passion for nutrition and its intricacies and hidden treasures is still very much alive, and I would like to share with you some of the valuable things I have learnt in the past 3 years of research.
I am tackling a debatable issue which needs no introduction: Glycaemic Index (GI). I leave it up to you to decide whether it warrants the hype and billion-dollar/rand industry it has been turned into.
The Glycaemic index of foods: a physiological basis for carbohydrate exchange. Jenkins, DA et al. (1981); Am J Clin Nutr:34. 362 – 366.
Jenkins and his team set out to measure the actual blood glucose response in man from different carbohydrate sources, since evidence showed that the information from exchange lists may not reflect the physiological effect of foods.
The study included 34 subjects (almost twice more males (n=21) than females (n=13). The average age was 29 years.
Fifty six foods were given in 50-g carbohydrate portions, after an over-night fast calculated from food composition tables from the Medical Research Council: London.
Palatability of meals was compensated for by including milky tea and 300 ml of milk with all breakfast cereals. GTT’s were taken over the same time as meals, with glucose as the comparative carbohydrate. Meals were eaten within 10 – 15 minutes. Finger prick samples were obtained 0, 15, 30, 45, 60, 90 and 120 minutes post-prandially.
Except for dairy products, there was great variation within most food groups. Although there was a rather large difference in the GI of different starchy foods, the GI seemed to flatten out above a carbohydrate consumption of 50 g. (See Figure 1). At an intake of 25 g of carbohydrate, very little difference was seen between white bread and glucose.
Peas, soya beans, peanuts and root vegetables (carrots) were difficult to consume during the allowed time, and in
the case of large volumes the carbohydrate content was reduced to 25 g.
Protein and fat was shown to have a negative relationship with GI. In their discussion, the researchers tried to explain why this phenomenon occurs. Fat retards gastric emptying, and protein elicits a slight insulin response, probably facilitating glucose clearance. Both dietary fibre and table sugar had no effect.
Response curve to glucose
No difference was found between ordinary pasta, wholemeal pasta, brown rice and their low fibre counterparts.
The authors commented: “One striking feature was that the high carbohydrate foods with the lowest glycaemic index were those eaten commonly by the poor in Western countries or the inhabitants of large parts of Africa and Asia.”
It surprises that no significant relationship was found between GI and dietary fibre. This might relate to the fact that high fibre foods originated from wheat, which has a fibre with little effect on blood glucose. Soluble fibre, on the other hand, might explain why the intake of legumes is known to flatten the blood glucose rise after 50 g carbohydrate intake more than other fibre analogues.
The study does mention that the high fructose content of some fruits might be accountable for it low GI (definition of GI: glucose response with an insulin effect – fructose does not apply)
The sugar content of food had no relation to blood glucose response, presumably due to the very small rise produced by fructose and reflected in the response to sucrose.
An insulin index of foods: the insulin demand generated by 1000 kJ portions of common foods. Holt, SHA, Brand Miller JC and Petocz, P. Am J Clin Nutr (1997); 66: 1264 – 1276.
More than a decade after the initial introduction to GI by Jenkins and his team, a team from Australia set out to measure the actual insulin index from different foods. Finger prick blood samples were taken similarly to the GI study, but white bread was used as the reference food. The sample size was slightly larger (n=41). From the area under the insulin curve an insulin score was calculated, using 1000 kJ (420 calories) as the portion size, as opposed to 50 g carbohydrate in the GI study.
Carbohydrate is not the only nutrient responsible for an insulin response. Protein added to carbohydrate can have a modest rise in insulin secretion, without raising blood glucose. This effect is of particular worth in treatment of diabetic patients. Adding fat to a carbohydrate-rich meal increases insulin secretion, even though plasma glucose levels are decreased. This shows that postprandial insulin responses are not always to the total carbohydrate content of the meal, or the postprandial glucose response.
Although proteins produced insulin responses, but in general less than starch-based foods, baked beans, if taken as a protein food illicit a significantly higher insulin response, as it contains much more carbohydrate than higher quality proteins.
In the fruit group, bananas and grapes had the highest insulin score, which can be explained through their higher glucose: fructose ratio than other fruits. In terms of total carbohydrate, these fruits are similar to other fruits