The vast majority of people with diabetes, however, don’t have this extreme form of the disease. Any dysfunction in the insulin-glucose system - all of which end up as the same basic problem - can be the cause: the body may not produce enough insulin; it may not respond to insulin properly; or it may produce too much insulin.

The classic way in which diabetes is believed to occur is that it starts as ‘insulin resistance’, where the body’s cells fail to react to insulin properly. This, in turn, causes the pancreas to produce more insulin to compensate, leading to a vicious circle that culminates in a kind of system exhaustion as the pancreas becomes overworked and ultimately stops functioning.

This kind of diabetes is called type 2, and is the sort that is increasing alarmingly not just among middle-aged adults, who are the typical sufferers, but in children as well.

If symptoms are noticed early enough, type 2 diabetics won’t need to be given artificial insulin as long as they strictly control their intake of carbohydrates. So, because of this type of diabetes, nutritionists have had to look at carbohydrates in a new light, which has led to a major rethink about the causes of diabetes.

What causes diabetes? Originally, nutritionists analysed foods by the amount of energy they provided - in other words, their calories. Calorie levels were worked out in the laboratory rather crudely by burning foods and seeing how much heat energy they produced.

However, caloric measures are useless for diabetics, who need to know the glucose value of foods to enable them to control their blood glucose levels through their diet, and so avoid the need for insulin injections. But the glucose values of foods can’t be measured in the laboratory; they must be tested on live human beings.

So, in a painstaking series of tests carried out in the 1980s, every foodstuff was analysed for its potential to produce glucose in the bloodstream. Human guinea pigs were fasted to create a baseline to measure against, then given a single food to eat. Regular blood samples were taken over four hours to chart the changes in blood glucose levels.

The results didn’t turn out quite as expected. One surprise was the finding that all carbohydrates cause a glucose peak roughly 30 minutes after ingestion. Previously, it had been thought that simple carbohydrates (like sugar and honey) were fast-acting while complex carbohydrates (like potatoes and cereals) were slow-acting.

Nevertheless, there were dramatic differences between carbohydrates in the levels of blood glucose they produced. And as the technical term for blood glucose is glycaemia, the differences between carbohydrates were measured in terms of a ‘glycaemic index’ (GI): the higher the number, the greater the amount of blood glucose produced by the food.

Nutritionists arbitrarily gave glucose itself a GI of 100 and, not surprisingly, virtually all foods are lower than that (see box, p 4). Also as expected, the most refined carbohydrates led to the highest blood glucose levels - but there were some surprises. Cooking was found to have major effects on the GI. Carrots, for example, produce three times the blood glucose level when cooked rather than eaten raw. In fact, any cooking or processing raises the GI significantly.

The GI has now been used by French scientist Dr Michel Montignac as the basis of a successful slimming regime - although, arguably, his greater contribution to health has been to develop a new theory of the causes of diabetes.