During the 1980s and 1990s, a number of controlled intervention studies in healthy individuals who maintained their body weight showed that high- carbohydrate diets often resulted in higher blood TG levels and lower HDL- cholesterol levels , changes that are atherogenic and increase the risk of coronary heart disease , despite improved total and LDL-cholesterol levels (29).
These findings sparked particular concern for people with diabetes because their lipid abnormalities tended to be higher TG and lower HDL- cholesterol level rather than the high total and LDL-cholesterol typically observed in non-diabetic individuals (18).
Hence the magnified risk of atherosclerosis in people with diabetes might be related to blood lipid risk factors that are specifically worsened by high-carbohydrate diets.
The biochemical mechanisms responsible for increased plasma TG levels following low-fat, high-carbohydrate diets remain uncertain but are clearly different to those responsible for elevated TG levels following increased fat intakes.
Parks et al. (30) demonstrated that high-carbohydrate diets reduce the clearance of VLDL-TG from the plasma, but do not increase VLDL-TG secretion or de novo lipogenesis in the liver as had been postulated.
The mechanisms by which high-carbohydrate diets decrease HDL- cholesterol are also unknown and should be a priority in future research.
In two recent cross-sectional studies of healthy adults, a significant inverse association was found between serum HDL-cholesterol concentration and dietary GI for both men and women (the higher the GI rating of the diet, the lower the HDL concentration) (31,32).
In fact, the glycaemic index of the diet was the only dietary variable significantly related to serum HDL-cholesterol.
These findings suggest that post-prandial glucose and insulin responses may directly influence HDL levels.
174 NUTRITIONAL MANAGEMENT OF DIABETES MELLITUS
THE SCIENTIFIC BASIS FOR RECOMMENDING HIGH-MUFA DIETS FOR DIABETES
Many diabetes experts argue in favour of allowing a higher MUFA intake for people with diabetes, on the grounds that high-carbohydrate diets can increase blood glucose, insulin and TG levels and reduce HDL-cholesterol levels.
A meta-analysis of nine studies with a total of 133 subjects comparing these two approaches to diet therapy in patients with diabetes revealed that high-MUFA diets (22, 33% of energy intake; total fat ¼ 37, 50% energy) improved lipoprotein profiles as well as glycaemic control (19).
Compared to high- carbohydrate diets (50, 60% energy intake), high-MUFA diets reduced fasting TG and VLDL-cholesterol levels by about 20% and caused a modest increase in HDL-cholesterol (4%) but had no effect on LDL-cholesterol.
There was no evidence that high-MUFA diets induced weight gain in these tightly controlled studies.
However, there are several limitations that need to be raised before deciding whether they provide sufficient evidence to formulate recommenda- tions for therapeutic diets: .
None of the studies controlled for/considered the confounding effects of the glycaemic index of the high-carbohydrate diets. .
The diets contained relatively small amounts of fibre (530 g/day, mostly in processed form). .
The studies were conducted under tightly controlled conditions, not allowing spontaneous weight loss/weight gain to occur. .
Most of the studies were of very short duration (two to four weeks), the longest being six weeks. .
A third of the studies were conducted by the same research group. .
Improvement in glycaemic control was assessed on the basis of urinary glucose and fasting, pre-prandial, post-prandial or 24-h blood glucose and insulin profiles. .
Notably, in the six studies that assessed HbA1c or fructosamine (the best markers of long-term glycaemic control), none of the changes were significant.
Thus we lack the evidence that high-MUFA diets improve overall diabetes control by the most valid measure of disease risk (i.e.
HbA1c).
This contrasts with the consistent effect of low-GI, high-fibre, carbohydrate-rich diets in lowering HbA1c (see below).
Furthermore, the positive effects of high-MUFA diets on blood lipids are often seen only when the high-MUFA diet is extremely high in fat (as much as 45, 50% of energy) and very low in carbohydrate (about 35% of energy) (33).
In studies with smaller and more realistic dietary changes, the effects of MUFA on blood lipids are more modest.
One can question the effect of such a very high-fat diet on insulin sensitivity, weight control and ability of patients to comply.
In fact, the largest study of this kind suggests that
THE ROLE OF CARBOHYDRATE 175
the beneficial effects of MUFA on insulin sensitivity disappear when fat intake exceeds 38% total energy (33). These studies confirm that there are definitive adverse risks associated with low-fibre, high-carbohydrate diets. However, they do not prove that the original recommendation to increase both carbohydrate and fibre was wrong, nor do they allow us to say whether a diet rich in monounsaturated fat is better than a high-fibre, high-carbohydrate diet.
THE EVIDENCE THAT LOW GLYCAEMIC INDEX, HIGH-CARBOHYDRATE DIETS ARE SUPERIOR
It is now well established that both the type and amount of carbohydrate influences the degree of post-prandial glycaemia (34).
The type of carbohydrate is best described by its glycaemic index, a ranking of foods according to their immediate effect on blood glucose levels (Figure 11.1).
Per gram of carbohydrate a food with a GI of 80 (e.g. potato) has twice the glycaemic impact of a food with a GI of 40 (e.g. pasta) and this applies even in mixed meals (35,36) (Figure 11.2).
The proportions of starch, sugar, fat or fibre in foods are not a good guide to GI.
Many common starchy foods (even wholemeal versions) such as bread, rice and breakfast cereals have surprisingly high GI values, while foods containing sugars often have a relatively low GI (37).
Reducing the overall GI of the diet involves substitutions within those food groups that contribute most of the dietary carbohydrate (Table 11.1).
The GI of foods is highly relevant to the management of Type 1 and Type 2 diabetes.
In nine well-designed long-term studies in diabetic subjects, low-GI diets (GI values 555%) were shown to reduce glycosylated proteins (HbA1c and/or fructosamine) by an average of almost 11% over periods ranging from two to 12 weeks (38).
At the end of the low-GI, high-carbohydrate diet, urinary C-peptide levels (a measure of endogenous insulin demand) fell by an average of 20%, daytime blood glucose levels decreased by 16%, and total cholesterol and TG were reduced by 6% and 9%, respectively (39).
Triglyceride levels fell to a much larger extent (by up to 20%) in patients with overt hypertriglyceridemia.
In a recent, randomised, cross-over study, clotting factors were normalised in patients with Type 2 diabetes by a low-GI, high-carbohydrate diet, but unchanged by a high-GI diet containing similar amounts of energy, protein, fat, carbohydrate, starch and fibre (40).
Studies comparing the effects of high-GI versus low-GI carbohydrate-rich diets have been longer (four to 12 weeks) than the high-MUFA studies, and unlike the latter, have been able to document beneficial changes in HbA1c and/or fructosamine levels.
In the few studies that have directly compared high-carbohydrate, low-GI diets with high-MUFA diets, HDL levels were increased on both (compared to the
176 NUTRITIONAL MANAGEMENT OF DIABETES MELLITUS
Figure 11.1 The derivation of the glycaemic index
high-carbohydrate, high-GI diet) but insulin secretory function and sensitivity appeared to be better only on the low-GI diet (41,42). Reductions in HbA1c or fructosamine levels of 10% on the low-GI diets have been criticised as being ‘modest’, yet changes of this magnitude are commonly seen with oral hypoglycaemic drugs. Furthermore, these ‘modest’ changes were achieved in patients in free-living conditions, not in a controlled metabolic ward situation. Although the high- and
Figure 11.2 The GI of single foods predicts the GI of mixed meals in subjects withType 2 diabetes Source: Redrawn from Bornet et al. (35).
THE ROLE OF CARBOHYDRATE 177
Table11.1 A low-GI diet is achieved by substituting high-GI foods with GI alternatives. Breads, breakfast cereals and potatoes contribute the majority of carbohydrate in Western diets. Changes within these food groups have the biggest impact on the diet’s overall GI
High-GI food Low-GI alternative
Bread, ordinary wholemeal Bread containing a high proportion of wholegrains or white (‘granary’ breads), sour dough breads, stone ground breads Most breakfast cereals Unrefined cereal such as oats (muesli or porridge). Some processed cereals (e.g. All-Bran) Potato (all varieties) Sweet potatoes, pasta, noodles, legumes Most varieties of rice Basmati or other high amylose rices Cakes, biscuits and muffins Versions made with fruit, oats, wholegrains Tropical fruits such as Temperate climate fruits such as apples and stone fruit bananas
low-GI diets were usually designed to be similar in macronutrient composition, in some studies the low-GI diet contained more fibre.
Indeed in one outpatient study, dietary education with emphasis on low-GI foods resulted in higher carbohydrate and fibre intakes and less saturated fat intake than achieved by patients given ‘traditional’ dietary counselling (43).
However, on the whole, studies comparing high- and low-GI diets have contained much less fibre (550 g per day) than the earlier studies that provided the basis for recommending high-carbohydrate diets in diabetes (475 g per day).
Lastly, low-GI, high-fibre diets may be the only strategy (diet or drug) that enables HbA1c to be improved while simultaneously reducing the incidence of hypoglycaemic episodes in Type 1 diabetes (44).
These findings suggest that any adverse effect of high-carbohydrate diets on blood lipids is almost certainly linked to the high GI of most such diets.
Indeed, any strategy that slows down the rate of digestion and absorption of carbohydrate (e.g. nibbling versus gorging, alpha-glucosidase therapy or purified supplements of viscous fibre, as well as low-GI diets) has been shown to improve glucose and lipid metabolism in diabetes (45).
WHICHDIET ISBEST FORIMPROVING INSULINSENSITIVITY?
The body’s sensitivity to the hormone insulin predicts how well it handles a meal containing carbohydrate, i.e. how easily and quickly it restores normal glucose levels after consumption. In insulin-resistant states, large amounts of insulin are needed to restore euglycaemia and glucose and/or insulin levels may still be high 2 h later. In Type 2 diabetes, insulin resistance is often severe and is combined with impairments in insulin secretory capacity. Obesity, particularly