by William R. Ware, PhD
Judging dietary modifications is complex. There are many potential endpoints to examine, including initial weight loss, long-term weight loss maintenance, changes in blood lipids, and changes in markers of insulin resistance and carbohydrate metabolism. The latter relate to the various aspects of diabetes and the prediabetic state. One can also look at overall mortality and the incidence, progression and mortality associated with specific diseases.
The diet literature is vast and contains a mixture of results ranging from nonsense to significant. This review will feature carbohydrate restriction, which is more or less synonymous with low-carbohydrate diets, and how this relates to diabetes, prediabetes and the metabolic syndrome. Heart disease is indirectly related through its association with diabetes and the metabolic syndrome.
In the Research Report Saturated Fat: Friend, Foe or Neutral?, it was pointed out that while this type of fat is considered sufficiently dangerous that many organizations recommend aggressively reducing its percentage contribution to total energy intake, the review presented extensive documentation for the assertion that this recommendation was not evidence-based and in fact may not be justified at all. But the recommendation to lower saturated fat is part of the low-fat movement that started 30-40 years ago and resulted in the food industry manufacturing a vast collection of low-fat foods which found a ready market among a frightened population. Then, as has been documented in the literature and in the articles and the comprehensive book Good Calories, Bad Calories by Gary Taubes, the search for evidence came up empty handed. Saturated fat and the other types of fat were neutral or beneficial. Even the notion that the omega-6 polyunsaturated fats were bad because they were inflammatory has now come under criticism with the American Heart Association now recommending not avoiding this particular type of fat. But the low-fat diet has been the centerpiece of much nutritional advice coming from mainstream medicine and the academic and professional nutritional community. They would now appear to be in a position where their advice is not as evidence-based as they would like us to believe, and already subtle changes are seen in recommendations; i.e. more flexibility and less emphasis on low-fat. But overall, it must be a hard pill to swallow. This is especially true now that it has become difficult to demonize the low-carbohydrate diet on the basis of its increased fat content, given that the fat is bad hypothesis now lacks that important imprimatur, solid evidence.
Strongly related to the above history is the rebirth of interest in carbohydrate restriction and the attempt to give the low-carbohydrate type diet a level of respectability that it heretofore has lacked. It is worth reminding readers that it was not that long ago when Dr. Robert Atkins was hauled up before a U.S. congressional committee and accused of being a public enemy for suggesting that low-carbohydrate diets were healthy and the way to lose weight and decrease the risk of heart disease and diabetes. While this attack by the government and the "fat is bad" community failed, there was the terrible press after his death which suggested that he died because he followed his diet, but this was not true. One of his strongest opponents was in fact an organization with a fancy and convincing title which was apparently a front for an animal rights group of vegetarians. Nevertheless, over the years, brave investigators have nibbled away at the credibility of the anti-low-carb community and created a much more balanced view of what might be called the macronutrient distribution problem in human nutrition.
The low-fat revolution resulted in motivating individuals to increase dramatically their carbohydrate intake, frequently with little regard for the type of carbohydrate involved. Low-fat foods were purchased with abandon in spite of the fact that many had high levels of sugar and refined carbohydrates. Some nutritional scientists present a catalogue of what they believe were unintended and very unfavourable results. These include strongly elevated triglycerides, depressed HDL levels, increased levels of small, dense atherogenic LDL particles, decreased insulin sensitivity, glucose intolerance, prediabetes and type 2 diabetes, elevated risk of cardiovascular disease, increased inflammation, hyperglycemia, and hyperinsulinemia and finally the so-called metabolic syndrome or syndrome-X. Added on to this remarkable list was a trend toward either being overweight or obese. While it is true that some of these adverse results could have been avoided by very carefully and knowledgably selecting carbohydrates used to replace fat or by weight loss, the general public was ill prepared to do this nor was the food industry seen to provide much help. Even "whole wheat bread" was not really whole grain bread, but many had no idea there was any difference. The supermarkets reached the point where a substantial fraction of what they sold and in fact are still selling would not have been recognized a century ago as food at all. The age of ersatz food had arrived.
The above state of affairs is no doubt responsible for the growing appearance in the literature of calls for a reconsideration of the low-carb or carbohydrate restricted approach to eating, with special reference to the obese, those with the dyslipidemia characterized by low HDL and high triglycerides, the prediabetics and those with type 2 diabetes. This was in general what was being recommended 10-20 years ago, not only by Atkins, but also by the other authors of low-carb diet books, and as well endocrinologists like Dr. Diana Schwarzbein and specialists in the treatment of diabetes such as Dr. Richard K. Bernstein. Even the Optimum Weight for Life program at Children's Hospital in Boston has moved strongly in this direction under the guidance of Dr. David Ludwig, who appears to view the basic philosophy as simply endocrinology 101.
Of special significance is the use of carbohydrate restriction for diabetics and prediabetics. The failure of the ACCORD trial prompted Westman and Vernon1 to ask "Has carbohydrate-restriction been forgotten as a treatment for diabetes mellitus?" They describe a typical approach to glucose control by diabetics called "cover the carbohydrate" where high carbohydrate diets are consumed and then intensive medication therapy used to attempt to achieve blood sugar control. They even comment that they frequently see individuals who are instructed to eat high carbohydrate diets and counteract the effects with injectable glucose lowering therapy, which they regard as an open invitation to serious hypoglycemia which can have permanent adverse effects including increased mortality. What is interesting about these observations is that diabetics are being advised, presumably by either their physicians or nutritionists, to consume diets high in carbohydrates which aggravate the swings in serum glucose. There are many issues here including increased risk of vascular damage, kidney damage, ocular damage, and limb amputation. ACCORD was halted after the intensive insulin arm of the study showed enhanced mortality in spite of achieving, to some extent, blood glucose control.
Westman and Vernon also quote from the famous Principles and Practice of Medicine, 9th Edition, by Osler and McCrae. William Osler became famous at McGill, Johns Hopkins and finally as Regius Professor at Oxford. The edition cited, coauthored by McCrae, was published in 1923, just after Osler's death in 1919 and only two years after the discovery of insulin. Osler was considered the high priest of early 20th century medicine. In this text, the treatment of diabetes consisted of a diet 75% fat, 17% protein, 6% alcohol and only 2% carbohydrate with a daily energy intake of about 1800 calories. This was about 9 g/day of carbohydrate which makes the modern carbohydrate restriction diets at 50-100 g/day appear very liberal. Taubes in his book Good Calories, Bad Calories, presents a detailed history of the treatment of both diabetes and obesity in this early period in the evolution of modern medicine and finds the same picture of severe carbohydrate restriction as the basic therapy employed. One presumes that in these much earlier times, carbohydrate restriction had been found to "work."
This review will examine carbohydrate restriction, especially with regard to prediabetes, type 2 diabetes, and the so called atherogenic dyslipidemia associated with the metabolic syndrome, a blood lipid picture that some considered consider vastly more alarming than elevated LDL. Readers of this Newsletter who are diabetic should worry about limiting adverse vascular effects which can be deadly and destroy the quality of life. Prediabetics should worry about reversing the trend toward diabetes. Those with the metabolic syndrome should worry about getting rid of this condition. And finally, everyone else needs to worry about maintaining normal glucose metabolism and low levels of inflammation. In discussing carbohydrate restriction, these are among the issues we will discuss.
Carbohydrate restriction implies a low-carbohydrate diet, although the amount of carbohydrate restriction may well exceed that of some low-carbohydrate diets which in fact involve a rather small reduction to yield a diet still relatively high in this macronutrient. Frequently, discussions of low-carbohydrate diets involve the percentage of energy intake represented by this macronutrient rather than the absolute number of grams. For example, an intake of < 200 g/day of carbohydrate has been termed by some a low-carbohydrate diet, whereas others think it should be defined as the range of 50-150 g/day. For an 1800-calorie diet, this would represent 11% to 33% of energy intake. However, levels below 33% are generally necessary to force the body to burn ketones, the so- called ketogenic diets.2 Low-carbohydrate diets almost always result in lower energy intake due to the impact on satiety and appetite from the higher percentage or absolute amounts of fat and protein.2
There are three principal issues: (a) the influence of carbohydrate restriction on serum markers related to or alleged to be related to cardiovascular risk, which should include the risk of atherosclerosis and its progression; (b) the impact of carbohydrate restriction on carbohydrate metabolism, fat metabolism and storage and insulin sensitivity; and (c) weight loss.
CARBOHYDRATE RESTRICTION AND CARDIOVASCULAR RISK FACTORS
With regard to the first issue, Krauss et al3 published a key paper in 2006. In a controlled diet experiment, after a one-week normalization period with a basal diet consisting of 54% of energy from carbohydrate, four diets were used. The distribution of macronutrients is shown in the table. During the next 3 weeks no attempt was made to control weight. Then a 1000-calorie reduction was instituted for 4 weeks to reduce weight followed by a four- week period of weight stabilization by calorie adjustment, while keeping the energy distributions constant. Data was not provided regarding the distribution of energy intake by macronutrient at baseline.
In a letter to the editor, Westman et al 5 congratulate Krauss et al for presenting one of the strongest cases to
date for dietary carbohydrate restriction and provide a number of references to other studies which found results
that were similar for one or more aspects studied by Krauss et al, i.e. a satisfactory level of consistency. They
also point out that given the difficulty in loosing weight, the data of Krauss et al support the notion that
carbohydrate restriction is the default diet for the treatment of atherogenic dyslipidemia. Most of the beneficial
effects were seen prior to the weight loss phase of the study.
Many diet studies examine the impact on LDL cholesterol. But in the context of primary prevention, non-invasive
coronary artery imaging suggests that there is no connection between LDL and coronary artery plaque burden or
progression and LDL does not drive atherosclerosis.6 In these imaging studies it was also observed that among
the various lipid risk factors, only elevated HDL and a low APO B/APO A-1 ratio were frequently found to have a
beneficial influence on the rate of progression. The carbohydrate restricted diet with or without high saturated fat
significantly changed these two markers in the right direction. In addition, the results of Krauss et al also fail
completely to support the notion that elevated intake of saturated fat is bad in the context of atherogenic
dyslipidemia. Quite the contrary, saturated fat appears to enhance the benefits of carbohydrate restriction.
Incidentally, the increase in fat that Krauss et al used for the high saturated fat diet compared to the low-fat diet
corresponds to the opposite of the decrease recommended by mainstream medicine and nutrition where a
reduction from 15% to 7% is advised.
Two recent randomized studies with significant reductions in energy from carbohydrates and energy intake
showed the expected pattern of increasing HDL and decreasing triglycerides. One study had a decreased over
12 months from 46 to 34% in energy from carbohydrates, an increase from 36 to 44% for fat, and an energy
decrease of about 300 cal. An increase was seen in HDL of about 5 mg/dL and a decrease in triglycerides of
almost 30 mg/dL with no significant change in LDL.7 In another study, carbohydrate intake decreased over 24
months from 51 to 41% of energy, fat increased from 32 to 39%, energy intake decreased by 550 cal. HDL
increased by 8.4 mg/dL and triglycerides decreased by 24% with no significant change in LDL. Westman et al 2
have reviewed earlier studies where low-fat diets with carbohydrate intake ranging from 51-62% of energy were
compared with low-carbohydrate diets with carbohydrate intake at 8 to 37%. The same general picture appears
where the low-carbohydrate diets resulted in large decreases in triglycerides, large increases in HDL, very small
changes in LDL and weight loss ranging from 5 to 12 kg. These diets had energy intakes between 1300 and
1800 cal. For the low-fat diets with energy intakes ranging from 1100 to 1600 cal, the changes in blood lipids
were much smaller as was the weight loss. These studies ran from 6 to 12 months. Thus the evidence appears
compelling that low-carbohydrate diets improve the blood lipid picture associated with dyslipidemia.
In a review calling for a critical reappraisal of carbohydrate restriction, Accurso et al8 summarize the situation as
of 2008. It was concluded that substitution of fat for carbohydrate is generally beneficial for markers for CVD and
these beneficial effects of carbohydrate restriction do not require weight loss.
When Atkins put forward the low-carbohydrate notion and his highly carbohydrate restricted diet, mainstream
medicine was up in arms claiming it was very dangerous. This was in keeping with the "fat is public enemy
number one" dogma and also consistent with a lack of much understanding concerning diet and cardiovascular
risk. It was simply assumed on the basis of weak and inconsistent evidence that increasing fat consumption to
compensate for the drop in carbohydrates would produce a significant increase in the risk of heart disease. At
this time the dogma reigned supreme and low-fat diets and low-fat foods were promoted by the so-called experts
in the absence of much if any evidence. But now there is a large body of evidence concerning this issue and the
time has come, as Accurso et al point out, for a dispassionate examination of what appears to be a better
alternative, no matter how distasteful it is to those who have based their whole careers on the "fat is bad"
In a letter to the editor, Westman et al 5 congratulate Krauss et al for presenting one of the strongest cases to date for dietary carbohydrate restriction and provide a number of references to other studies which found results that were similar for one or more aspects studied by Krauss et al, i.e. a satisfactory level of consistency. They also point out that given the difficulty in loosing weight, the data of Krauss et al support the notion that carbohydrate restriction is the default diet for the treatment of atherogenic dyslipidemia. Most of the beneficial effects were seen prior to the weight loss phase of the study.
Many diet studies examine the impact on LDL cholesterol. But in the context of primary prevention, non-invasive coronary artery imaging suggests that there is no connection between LDL and coronary artery plaque burden or progression and LDL does not drive atherosclerosis.6 In these imaging studies it was also observed that among the various lipid risk factors, only elevated HDL and a low APO B/APO A-1 ratio were frequently found to have a beneficial influence on the rate of progression. The carbohydrate restricted diet with or without high saturated fat significantly changed these two markers in the right direction. In addition, the results of Krauss et al also fail completely to support the notion that elevated intake of saturated fat is bad in the context of atherogenic dyslipidemia. Quite the contrary, saturated fat appears to enhance the benefits of carbohydrate restriction. Incidentally, the increase in fat that Krauss et al used for the high saturated fat diet compared to the low-fat diet corresponds to the opposite of the decrease recommended by mainstream medicine and nutrition where a reduction from 15% to 7% is advised.
Two recent randomized studies with significant reductions in energy from carbohydrates and energy intake showed the expected pattern of increasing HDL and decreasing triglycerides. One study had a decreased over 12 months from 46 to 34% in energy from carbohydrates, an increase from 36 to 44% for fat, and an energy decrease of about 300 cal. An increase was seen in HDL of about 5 mg/dL and a decrease in triglycerides of almost 30 mg/dL with no significant change in LDL.7 In another study, carbohydrate intake decreased over 24 months from 51 to 41% of energy, fat increased from 32 to 39%, energy intake decreased by 550 cal. HDL increased by 8.4 mg/dL and triglycerides decreased by 24% with no significant change in LDL. Westman et al 2 have reviewed earlier studies where low-fat diets with carbohydrate intake ranging from 51-62% of energy were compared with low-carbohydrate diets with carbohydrate intake at 8 to 37%. The same general picture appears where the low-carbohydrate diets resulted in large decreases in triglycerides, large increases in HDL, very small changes in LDL and weight loss ranging from 5 to 12 kg. These diets had energy intakes between 1300 and 1800 cal. For the low-fat diets with energy intakes ranging from 1100 to 1600 cal, the changes in blood lipids were much smaller as was the weight loss. These studies ran from 6 to 12 months. Thus the evidence appears compelling that low-carbohydrate diets improve the blood lipid picture associated with dyslipidemia.
In a review calling for a critical reappraisal of carbohydrate restriction, Accurso et al8 summarize the situation as of 2008. It was concluded that substitution of fat for carbohydrate is generally beneficial for markers for CVD and these beneficial effects of carbohydrate restriction do not require weight loss.
When Atkins put forward the low-carbohydrate notion and his highly carbohydrate restricted diet, mainstream medicine was up in arms claiming it was very dangerous. This was in keeping with the "fat is public enemy number one" dogma and also consistent with a lack of much understanding concerning diet and cardiovascular risk. It was simply assumed on the basis of weak and inconsistent evidence that increasing fat consumption to compensate for the drop in carbohydrates would produce a significant increase in the risk of heart disease. At this time the dogma reigned supreme and low-fat diets and low-fat foods were promoted by the so-called experts in the absence of much if any evidence. But now there is a large body of evidence concerning this issue and the time has come, as Accurso et al point out, for a dispassionate examination of what appears to be a better alternative, no matter how distasteful it is to those who have based their whole careers on the "fat is bad" dogma.
CARBOHYDRATE RESTRICTION, CARBOHYDRATE METABOLISM AND DIABETES
The second issue associated with carbohydrate restriction and low-carbohydrate diets concerns carbohydrate metabolism, insulin sensitivity (or resistance, the other side of that coin) and fatty acid metabolism, all of which concern aspects of the metabolic syndrome. As carbohydrate metabolism becomes dysfunctional, first the prediabetic state and then diabetes occur. A low-fat diet (< 7% of energy) in combination with regular exercise is the current recommendation for preventing or treating diabetes.9 This appears counter intuitive considering that dietary carbohydrate is the major determinant of post-meal blood glucose levels. The American Diabetes Association, in its latest dietary guidelines, even suggest that if table sugar is added to the meal plan, all one needs to do is cover its impact with insulin or other glucose-lowering medications.9
Westman et al recently discussed the impact of insulin on the treatment of diabetes in the early 20th century.10 As mentioned above, a very low carbohydrate, high fat diet was employed for the treatment of type 1 diabetes using urine glucose levels as an indicator of efficacy, and this was the approach recommended in the principal medical text book of the time. After insulin was discovered, insulin injections were viewed as the ultimate solution, but studies consistently show that diabetics achieve poor blood sugar control and adverse vascular effects are widespread. In recent times, one approach has been to attempt to modify the impact of a high carbohydrate intake by emphasizing low glycemic index foods. A recent meta-analysis of a number of studies on both type 1 and type 2 diabetics suggests that this approach has a very limited impact on glycated hemoglobin A1c levels, the standard measure of long-term glucose control.11 However, as part of the Cardiovascular Health Study, Mozaffarian et al12 have examined the combined impact of lifestyle factors which included both low glycemic foods and exercise on the incidence of diabetes in later life in lean individuals and found dramatic decreases in the risk of developing diabetes, see Saturated Fat: Friend, Foe or Neutral?
As discussed in an earlier Research Report on the diagnosis of diabetes Diabetes and Pre-Diabetes, fasting glucose, the glucose tolerance test, and the blood levels of glycated hemoglobin A1c (HbA1c) are all used to access the short and long term status of glucose metabolism. As Accurso et al8 discuss in their recent review, carbohydrate restriction slows glycemic (blood sugar) responses and insulin response. In one study of obese diabetics, 14 days on a low-carbohydrate diet with an intake of about 1000 cal resulted in a drop in HbA1c (7.3% to 6.8%), and insulin sensitivity increased by 75%. In another study of a similar group, a very low-carbohydrate diet reduced blood glucose levels to normal over 48 weeks. The authors cite other studies that had similar if not greater impact on both HbA1c and glycemic control and even moderate carbohydrate reduction was reported to improve glycemic control by 40-55%. Carbohydrate restriction has been reported by several investigators to cause diabetic patients to reduce or eliminate medication. Feinman and Volek13 point out that the reduction or elimination of glycemic control medication is a pre-requisite for establishing the merits of the low-carbohydrate approach in the treatment of diabetes. They cite three studies, as of 2008, which in fact demonstrate that this is what happens. Indeed, carbohydrate restriction is so effective that when combined with glycemic control drugs it can result in dangerous hypoglycemia. Finally, glycemic control has been shown to dramatically reduce the levels and excursions of both insulin and glucose over a 24-hour period as compared to the usual high-carbohydrate diet diabetics are apparently encouraged to eat, with these very favourable results seen after only 2 weeks on a low-carbohydrate diet.8 These beneficial effects directly impact prediabetics as well and can result in normal glucose metabolism and thus dramatically reduce the risk of developing diabetes. Also, in both the prediabetic and diabetic, these changes in glucose metabolism reduce the risk of vascular complication, which is really the name of the game.
A retrospective case study published in 2003 illustrates the power of carbohydrate restriction in this context.14 Fourteen patients with diagnosed diabetes (13 type 2, 57% female, age 35-52) were counseled to reduce their carbohydrate intake to 20 g/day. Once glycemic control was achieved 5 g per day of carbohydrate was added until urinary ketones were no longer detectable. This fixed the final level of carbohydrate intake. Oral hypoglycemic agents were discontinued at the start of the intervention, and insulin used only if necessary to keep blood glucose in the 150-200 mg/dL range. The median follow-up was 8 months. Ten of the 14 had an initial HbA1c which averaged 10.9% with a range of 16.8-9.5%, i.e. all very high, typical of diabetics and dangerous in terms of vascular damage. The dietary intervention reduced the HbA1c in all 10 to less than 6.5% (average 5.5%, range 4.7-6.3%). An HbA1c of 6.5% is the latest suggested cut-off for the diagnosis of diabetes and near the threshold for the onset of adverse vascular effect of hyperglycemia. Another patient went from 12.0 to 6.8% in 2 months, and one started at 12.7 and reached 7.6% in 13 months. Only 2 patients out of 14 failed to respond, and 10/14 no longer had an HbA1c diagnostic of diabetes. These changes occurred in some patients with minimal weight loss and all were either obese or morbidly obese. Triglycerides dropped on average by 50.3% and HDL increased. While it is recognized that one study which was not randomized is of limited value, these results are similar to other studies and provided the motivation for Westman and Vernon to suggest that carbohydrate restriction was superior to insulin therapy and is in fact the protocol they use clinically with success.1
Richard K. Bernstein, M.D., in his book Diabetes Solution (Little Brown, New York, 2003), also promotes carbohydrate restriction, regards the target HbA1c levels recommended by the various diabetes associations as being absurdly high, and aims in his clinic to achieve values well below that indicative of diabetes. This book is highly recommended for anyone with diabetes.
As part of the now famous Nurses' Health Study, it has been demonstrated that diets rich in vegetable sources of protein and fat may modestly reduce the risk of diabetes. The authors comment that in a previous investigation of the same cohort, it was found that a score reflecting a diet high in fat and protein and low in carbohydrates was not associated with increased risk of coronary heart disease in women.15
CARBOHYDRATE RESTRICTION AND WEIGHT LOSS
Diet and weight loss constitute a complex subject. Issues include starting weight, fat distribution, energy derived from each macronutrient, presence of concomitant exercise and its intensity, duration of the study, and gender. Adherence to an assigned diet is a big issue. Benefits aside from weight loss are generally measured by markers of perceived cardiovascular and metabolic risk and inflammatory markers. The matter is complicated by the fact that 20-30% of obese individuals are what is called metabolically healthy, i.e. they do not differ significantly in a variety of metabolic markers from healthy non-obese individuals. Other issues include presence or absence of diabetes and menopausal state.
Problems can be found with most studies. These involve duration which is too short to be very meaningful, small or very small numbers of participants, failure to document adherence, and comparison of diets that do not involve meaningful differences in a given macronutrient. With carbohydrate restricted diet studies, one finds a number that employed unrealistically low-carbohydrate intake such that long term maintenance of the diet would probably be impossible and extrapolating the results to more realistic but still quite low-carbohydrate intake is unclear. Other studies claiming to compare low- and high-carbohydrate diets actually end up comparing two relatively high-carbohydrate diets.
A common characteristic of diet studies is that during the first few months the mean weight change is 4-8 kg, but by the end of a year it approaches the baseline mean value. When two or more distinctly different diets are compared, a difference in weight loss may be seen which is generally greater during the first few months and the diets then tend to converge in the long term to a smaller and similar weight loss. One rarely sees a diet study where the weigh continuously declines throughout the study. But there are also the extremes not reflected in the averages - some find it easy to continuously lose weight and some find it almost impossible. If one had to generalize, it is probably necessary to conclude that losing significant amounts of weight is difficult for many individuals.
The latest large randomized weight reduction trial to date reported in February 2009 and was heralded by the media as proving that all diets have the same effect on long-term weight loss regardless of their macronutrient composition, including high carbohydrate low-fat and low-fat high-carbohydrate.16 The diets were designed to be low-fat and average-protein, low-fat and high-protein, high-fat and average-protein, and high-fat and high- protein. In these 4 diets, the carbohydrate intake target as a percentage of total energy was 65%, 55%, 45% and 35%, respectively. However, it turned out that carbohydrate, fat and protein targets were never met and all four diets were in fact very similar during the execution of the study. Thus it would be a mistake to attach any significance to the weight-loss aspect of this study. While a number of other parameters were measured, when everyone is actually on the close to the same diet, the results are also not interesting. Some would no doubt argue that this study did not merit publication because the execution proved seriously flawed.
A second randomized trial that compared low-carbohydrate, Mediterranean and low-fat diets over 2 years has also recently reported.17 In terms of weight loss it was found that the Mediterranean and low-carbohydrate diets were effective alternatives to the low-fat diets. In the low-carbohydrate diet, carbohydrates were only decreased from 51 to 41%, but overall energy intake was decreased by 550 cal, whereas in the Mediterranean diet there was no significant change in the percentage of energy from carbohydrates and the energy decrease was about 370 cal. Both produced a weight loss at 24 months of about 6 kg whereas the low-fat diet with a decrease in energy of 570 cal and a constant percentage of energy from carbohydrates yielded a weight decrease of only 3 kg at 24 months.
A number of other studies could be cited but for the most part they were of short duration or involved unrealistically low-carbohydrate intake which could not be part of a long-term diet. Nevertheless, these short- term studies with severe carbohydrate restriction produced, as would be expected, large decreases in triglycerides and large increases in HDL. Thus the overall picture is consistent and suggests carbohydrate restriction offers a satisfactory protocol for modest weight loss and has an excellent impact with regard the atherogenic dyslipidemia which may enable this intervention to play a significant role in the prevention of cardiovascular disease. It is unfortunate that Robert Atkins did not live to see his ideas repeatedly confirmed.
CARBOHYDRATE RESTRICTION AND THE METABOLIC SYNDROME
Carbohydrate restriction improves the features of the metabolic syndrome and it has been suggested by Volek and Feinman that in fact the syndrome could be defined by the response to carbohydrate restriction.18 This point of view is of considerable interest since the metabolic syndrome carries a predisposition to diabetes, cardiovascular disease and other pathologic states. However, most formal guidelines and clinical papers have not emphasized carbohydrate restriction as a recommended approach in treating either the syndrome or its individual components. This may be due to a fear of conflict with the fat is bad dogma.
The usual definition of the metabolic syndrome used in North America is that any three of the following criteria must be met: (1) waist circumference > 102 cm in men and 88 cm in women; (b) triglycerides equal to greater than 150 mg/dL (1.7 mmol/L); HDL cholesterol < 40 mg/dL in men and < 50 mg/dL (1.0 and 1.3 mmol/L); blood pressure equal to greater than 130/85 and blood glucose > 110 mg/dL (6.1 mmol/L). Volek and Feinman18 present 13 carbohydrate restricted diet studies that favourably impacted triglycerides, HDL and glucose in both normal weight and obese men and women, as well as reducing weight and in some studies, systolic blood pressure. They also demonstrate on the basis of three studies that for short term low-fat vs. ketogenic (very low carb) diets, only the latter result in favourable changes in the lipid parameters. More recent studies discussed above support this observation.
In a recently published paper, Volek, Feinman and coworkers reported on a detailed study over 12 weeks of a carbohydrate restricted diet vs. a low fat diet.19 The study involved 40 subjects with elevated triglycerides and low HDL. Carbohydrate intake in the restricted diet was individually determined by the intake that resulted in a low level of ketosis. Foods consumed included unlimited amounts of beef, poultry, fish, eggs, oils and heavy cream, moderate amounts of hard cheeses, low-carbohydrate vegetables and salad dressings, and small amounts of nuts and seeds. The low fat diet was designed to provide < 10% of calories from saturated fat and< 300 mg of cholesterol. Foods encouraged included whole grains, fruit, vegetables, vegetable oils, low fat and lean meat. The results demonstrated that a diet restricted in carbohydrates can provide a more comprehensive improvement in the clinical risk factors associated with the metabolic syndrome than a low-fat diet at reduced caloric intake. Furthermore, the carbohydrate restricted diet showed more favourable responses to alternative indicators of cardiovascular risk such as post-meal lipidemia, the Apo B/Apo-A1 ratio, and the LDL particle size distribution, i.e. a decrease in the small dense LDL particles viewed by some a the atherogenic component of the total LDL. The carbohydrate diet also improved glycemic and insulin control.
There is growing recognition that atherogenic dyslipidemia presents an independent risk factor for coronary heart disease and may account for the large number of individuals with normal or low LDL that experience heart attacks. The connection is probably stronger with events than the progression of atherosclerosis where the evidence that elevated triglycerides and low HDL are associated with risk is weak but still suggestive.6 It is interesting to look at recent calls for action and intervention.20,21 One group uses the phrase "Residual Risk Reduction Initiative." These mostly discuss pharmaceutical interventions using fibrates or niacin or the increased consumption of long-chain omega-3 fatty acids such as are derived from marine sources. But when dietary interventions are discussed, no mention is made of carbohydrate restriction and the major study discussed is one which reduced carbohydrates only from 58 to 48% of total energy.22 Yet the changes in HDL and triglycerides that result from a diet more restricted in carbohydrates are comparable to what can be achieved with fibrates and/or niacin.23
There are two potential explanations: (a) the call for action is coming from medical scientists who have strong links with the pharmaceutical industry and think in terms of drug interventions, and (b) carbohydrate-restricted diets have not made it to mainstream medicine and, because they generally involve an increase in fat and even saturated fat, go against the fat is bad dogma which is still very much alive and well as suggested by the constant refrain that one must reduce fat intake. The trouble is that many individuals who qualify for treatment of atherogenic dyslipidemia or the metabolic syndrome will receive a long term drug program recommendation and not even be made aware of the merits of carbohydrate restriction. If dietary recommendations involve only the reduction of saturated fat from 15% to 7% of energy, restricting dietary cholesterol and reducing fat in general, then the dietary approach will probably fail and the recommendation of a pharmaceutical approach will gain credibility as the only answer. This is especially true if a large weight loss is also used as the target. This is unfortunate since it is not the only answer. Furthermore, while most diets produce only modest declines in weight after one or two years, the carbohydrate restricted diet can produce dramatic reductions in dyslipidemia with these modest weight loses an in fact eliminate the diagnosis of the metabolic syndrome.
EGGS AND THE CARBOHYDRATE RESTRICTED DIET
With the advent of the fat is bad era and the belief that dietary cholesterol was also bad, the nutritional and medical establishments took a very dim view of the egg, since the typical egg yolk contains 200 mg of cholesterol. This appears to lack justification given that for 70% of the population, dietary cholesterol does not significantly elevate serum cholesterol and for those responders, dietary cholesterol elevates both LDL and HDL and the ratio remains constant suggesting no increase, according to the conventional wisdom, in CHD risk. Also, if one does not believe that fat is bad, then the fat in eggs is also not an issue. Furthermore, research does not support a consistent relation between egg intake and increased CHD incidence and egg consumption has been shown to promote the formation of large LDL particles thought to be much less atherogenic.24
Two recent papers report results on the impact of egg consumption on lipid parameters and inflammation in subjects on carbohydrate-restricted diets. The first study found that adult men on a carbohydrate-restricted diet who consumed eggs had a significant increase in HDL compared to those consuming an egg substitute. The former had a 12% increase in HDL compared to the latter who experienced a 1.2% decrease, both without a significant change in LDL levels. Mean HDL levels went from 47.6 to 57.1 mg/dL after 12 weeks on the diet containing eggs. This diet provided an additional 640 mg/day of cholesterol. TG levels were significantly decreased in both groups and the addition of eggs to the diet did not alter the positive effects of carbohydrate restriction on features of the metabolic syndrome.25 A second study from the same group found that adding eggs to the carbohydrate-restricted diet made a significant contribution to the anti-inflammatory effects of the diet, perhaps due to the lutein in eggs.26
Current mainstream recommendations call for reducing dietary cholesterol. These results suggest exactly the opposite, at least in the context of a low-carbohydrate diet. Furthermore, eggs contribute to healthy diets, providing protein, dietary carotenoids, lecithin, lutein, zeaxanthin and choline. As Herron et al point out, eggs are particularly appropriate in elderly populations, especially since elevated total cholesterol appears to be a risk factor only in middle-aged individuals.27
In the context of the prevention of coronary heart disease and the prevention and treatment of diabetes as well as the prediabetic state, carbohydrate restriction would appear to the approach of choice, especially if the desire is to avoid pharmaceutical intervention if at all possible. Furthermore, carbohydrate restriction can beneficially impact both blood lipids and glucose metabolism even in the absence of significant weight loss, whereas even the smaller benefits achievable from the low-fat diet generally require weight loss. Furthermore, carbohydrate restriction has the potential to reverse prediabetes and diabetes and eliminate the need for glucose lowering medication. The fear that carbohydrate-restricted diets would have an adverse impact on the blood lipid profile does not appear evidence-based. Quite the contrary, this type of diet has the potential to reverse the dyslipidemia of the metabolic syndrome and its impact on LDL, if one believes that to be significant, is minimal. The severity of carbohydrate restriction depends on the goal of the intervention. Fortunately, most of the parameters involved are easily measured. If on is trying to reverse diabetes or prediabetes, then HbA1c and fasting glucose are of interest. Both are routine and one can be done at home. Assessing the metabolic syndrome involves only one measurement that cannot be done at home, the blood lipid profile, but this again is a routine measurement although now the focus is on triglycerides and HDL, not on LDL and total cholesterol. Thus if one is armed with a scale, a tape measure, a blood glucose meter and a blood pressure measuring device, aside from two blood tests, the success of a dietary intervention can be followed at home.