January 2015
Richard D Feinman, Wendy K Pogozelski, Arne Astrup, Richard K Bernstein, Eugene J Fine, Eric C Westman, Anthony Accurso, Lynda Frassetto, Barbara A Gower, Samy I McFarlane, Jörgen Vesti Nielsen, Thure Krarup, Laura Saslow, Karl S Roth, Mary C Vernon, Jeff S Volek, Gilbert B Wilshire, Annika Dahlqvist, Ralf Sundberg, Ann Childers, Katharine Morrison, Anssi H Manninen, Hussain M Dashti, Richard J Wood, Jay Wortman, Nicolai Worm 

 

Highlights

• We present major evidence for low-carbohydrate diets as first approach for diabetes.

• Such diets reliably reduce high blood glucose, the most salient feature of diabetes.

• Benefits do not require weight loss although nothing is better for weight reduction.

• Carbohydrate-restricted diets reduce or eliminate need for medication.

• There are no side effects comparable with those seen in intensive pharmacologic treatment.

 

Abstract

The inability of current recommendations to control the epidemic of diabetes, the specific failure of the prevailing low-fat diets to improve obesity, cardiovascular risk, or general health and the persistent reports of some serious side effects of commonly prescribed diabetic medications, in combination with the continued success of low-carbohydrate diets in the treatment of diabetes and metabolic syndrome without significant side effects, point to the need for a reappraisal of dietary guidelines. The benefits of carbohydrate restriction in diabetes are immediate and well documented. Concerns about the efficacy and safety are long term and conjectural rather than data driven. Dietary carbohydrate restriction reliably reduces high blood glucose, does not require weight loss (although is still best for weight loss), and leads to the reduction or elimination of medication. It has never shown side effects comparable with those seen in many drugs. Here we present 12 points of evidence supporting the use of low-carbohydrate diets as the first approach to treating type 2 diabetes and as the most effective adjunct to pharmacology in type 1. They represent the best-documented, least controversial results. The insistence on long-term randomized controlled trials as the only kind of data that will be accepted is without precedent in science. The seriousness of diabetes requires that we evaluate all of the evidence that is available. The 12 points are sufficiently compelling that we feel that the burden of proof rests with those who are opposed.

 

Introduction

Reduction in dietary carbohydrate as a therapy for diabetes has a checkered history. Before and, to a large extent, after the discovery of insulin, it was the preferred therapeutic approach [2]. Only total reduction in energy intake was comparable as an effective dietary intervention. The rationale was that both type 1 and type 2 diabetes represent disruptions in carbohydrate metabolism. The most salient feature of both diseases is hyperglycemia and the intuitive idea that reducing carbohydrate would ameliorate this symptom is borne out by experiment with no significant exceptions. Two factors probably contributed to changes in the standard approach. The ascendancy of the low-fat paradigm meant that the fat that would replace the carbohydrate that was removed was now perceived as a greater threat, admittedly long term, than the immediate benefit from improvement in glycemia. The discovery of insulin may have also cast diabetes—at least type 1—as a hormone-deficiency disease where insulin (or more recent drugs) were assumed to be a given and dietary considerations were secondary. For these and other reasons, dietary carbohydrate holds an ambiguous position as a therapy.

Although low-carbohydrate diets are still controversial, they have continued to demonstrate effectiveness with little risk and good compliance. At the same time, the general failure of the low-fat paradigm to meet expectations, coupled with continuing reports of side effects of different drugs, indicates a need for reevaluation of the role for reduction in carbohydrate. The current issue seems to be whether we must wait for a long-term randomized controlled trial (RCT) or whether we should evaluate all the relevant information. Practical considerations make it virtually impossible to fund a large study of nontraditional approaches. In any case, the idea that there is one kind of evidence to evaluate every scientific question is unknown in any science. Here we present 12 points of evidence supporting the use of low-carbohydrate diets as the first approach to treating type 2 diabetes and as the most effective adjunct to pharmacology in type 1. They are proposed as the most well-established, least controversial results. It is not known who decides what constitutes evidence-based medicine but we feel that these points are sufficiently strong that the burden of proof rests on critics. The points are, in any case, intended to serve as the basis for improved communication on this topic among researchers in the field, the medical community, and the organizations creating dietary guidelines. The severity of the diabetes epidemic warrants careful and renewed consideration of our assumptions about the diet for diabetes.

Definitions

A lack of agreed on definitions for low-carbohydrate diet has been a persistent barrier to communication. We propose the definitions in Table 1 to eliminate ambiguity. Each definition is based on use in multiple publications by those authors who have performed the experimental studies [3], [4], [5], [6].

Very low-carbohydrate ketogenic diet (VLCKD)
• Carbohydrate, 20–50 g/d or <10% of the 2000 kcal/d diet, whether or not ketosis occurs. Derived from levels of carbohydrate required to induce ketosis in most people.

• Recommended early phase (“induction”) of popular diets such as Atkins Diet or Protein Power.

Low-carbohydrate diet: <130 g/d or <26% total energy
• The ADA definition of 130 g/d as its recommended minimum.

Moderate-Carbohydrate Diet: 26%–45%
• Upper limit, approximate carbohydrate intake before the obesity epidemic (43%).

High-Carbohydrate Diet: >45%
• Recommended target on ADA websites.

• The 2010 Dietary Guidelines for Americans recommends 45%–65% carbohydrate. The average American diet is estimated to be ∼49% carbohydrate.

• Carbohydrate Consumption (NHANES)†:
• Men
• 1971–1974: 42% (∼250 g for 2450 kcal/d)

• 1999–2000: 49% (∼330 g for 2600 kcal/d)

• Women
• 1971–1974: 45% (∼150 g for 1550 kcal/d)

• 1999–2000: 52% (∼230 g for 1900 kcal/d)

ADA, American Diabetes Association; NHANES, National Health and Nutrition Examination Survey

∗ Derived from Accurso et al. [3] and references therein.

NHANES is a series of studies conducted since 1960 that monitors >5000 people.

We recognize that levels of carbohydrate tolerance vary between individuals and even in one person over time. For example, a very low-carbohydrate ketogenic diet (VLCKD) is defined as comprised of 20 to 50 g/d carbohydrate, but because of individual variability, ketosis (blood ketone bodies >0.5 mM) may not occur.

 

12 Points of evidence

Point 1. Hyperglycemia is the most salient feature of diabetes. Dietary carbohydrate restriction has the greatest effect on decreasing blood glucose levels

Both type 1 and type 2 diabetes are defects in the response to food, particularly to carbohydrates. The associated hyperglycemia is both the most characteristic symptom and the cause of downstream sequelae including insulin effects and generation of advanced glycation end products (AGEs). The most obvious glycation product, hemoglobin A1c (HbA1c) is widely taken as diagnostic. Glycemic control remains the primary target of therapy in patients with type 1 and type 2 diabetes. It is universally accepted that dietary carbohydrate is the main dietary determinant of blood glucose [7] and restriction shows the greatest reduction in postprandial and overall glucose concentrations as well as HbA1c [3], [6], [8], [9], [10], [11], [12], [13], [14]. Whereas defects in repression of gluconeogenesis and glycogenolysis are the major causes of hyperglycemia [8], [15], carbohydrate is by far the greatest dietary contributor to blood sugar rises and, as expected, dietary carbohydrate restriction reliably reduces glucose profile.

Hussain et al. [14], for example, compared a VLCKD with a low-calorie diet over a 24-wk period in 102 diabetic and 261 nondiabetic individuals. As shown in Figure 1, blood glucose dropped more dramatically in the VLCKD group than in those given the low-calorie diet. In the patients with type 2 diabetes, however, after 24 wk, the average blood glucose level was approximately 1 mM lower than in the low-calorie diet group. More significantly, the VLCKD group approached normal blood sugar levels after 24 wk, whereas the low-calorie group's blood glucose concentration leveled out at 16 wk and remained elevated. In the normal patients, blood glucose was already at normal levels, and the VLCKD produced only a small effect.

The second panel in Figure 1 shows the effect of the two diets on HbA1c levels. At 24 wk, patients with diabetes given the VLCKD achieved an HbA1c of 6.2%, whereas the average HbA1c in the low-calorie diet group remained >7.5%.

Point 2. During the epidemics of obesity and type 2 diabetes, caloric increases have been due almost entirely to increased carbohydrates

Data from the National Health and Nutrition Examination Surveys (NHANES) [16] indicate a large increase in carbohydrates as the major contributor to caloric excess in the United States from 1974 to 2000 (Fig. 2). From the time of the first NHANES study (1974) to the last (2000), dietary carbohydrate in men rose from 42% to 49% of calories. For women, carbohydrate rose from 45% to 52%. The absolute amount of fat decreased for men during this period and showed only a slight increase for women. The inset to the Figure 2 reveals the rise, during this period, of the incidence of type 2 diabetes to its current near epidemic proportions [17]. More recently, one study [18] analyzed U.S. Department of Agriculture availability data and found that the absolute fat availability had increased slightly, but, as shown in the NHANES data [19], the increase in carbohydrate was the predominant change.

 

Discussion

The need for a reappraisal of dietary recommendations stems from the following:

1. General failure to halt the epidemic of diabetes under current guidelines.

2. The specific failure of low-fat diets to improve obesity, cardiovascular risk, or general health (points 1 and 4).

3. Constant reports of side effects of commonly prescribed diabetic medications, some quite serious (points 12).

4. Most importantly, the continued success of low-carbohydrate diets to meet the challenges of improvement in the features of diabetes and metabolic syndrome in the absence of side effects.

The benefits of carbohydrate restriction are immediate and well documented. Concerns about the efficacy and safety of carbohydrate restriction are long term and conjectural rather than data driven. Most objections stem from the proposed dangers of total or saturated fat embodied in the so-called diet–heart hypothesis. At this point, the diet–heart hypothesis has had a record of very limited clinical or experimental success to support its position. The issue has become the subject of strong reaction in both the scientific literature and the popular press [48], [50], [51], [90] (point 8).

It is well established that weight loss, by any method, is beneficial for individuals with diabetes. The advantages to a low-carbohydrate approach are that, because of greater satiety, explicit calorie reduction on the part of the patient may not be required. There may be de facto reduction in calories without the need for replacement. The extent to which there is replacement, either fat or protein may be beneficial (points 4 and 6) although, in practice, fat is recommended unless there is already lower protein. Concerns about high protein in carbohydrate restriction have been raised but, except for those people with existing kidney disease, none has ever been demonstrated [91]. Protein also tends to a stable self-limiting part of the diet. Perhaps most important, if carbohydrate is low, glycemic control and other physiologic parameters are improved even if weight loss is not accomplished (point 3).

Finally, it should be recognized that the use of low-carbohydrate diets is not a recent experiment and may well approximate the diet used by much of humanity for tens of thousands of years before the rise of agriculture. Current knowledge dictates that carbohydrate restriction should be a default treatment for type 2 diabetes and a default adjunct therapy for type 1. Given the superior outcomes of carbohydrate-restricted diets, patients should not be discouraged from adhering to them as is frequently observed. They should, in fact, be encouraged to follow this approach.

The 12 points of evidence represent the best investigated and least conjectural ideas on diabetes. It is unlikely that one dietary strategy, any more than one kind of pharmacologic treatment will be best for all individuals. Patients can refuse medication or opt out of surgery, but they cannot not be on a diet and low-carbohydrate is the reasonable place to start. We recognize that there are many complications and issues that are still not understood, however, we have tried to isolate the factors that have the fewest contradictions.

This review emphasized the most obvious principles. An anonymous reviewer, however, raised two important if more conjectural points. We were asked “To specify role of starch versus mono- and disaccharides in carbohydrate-semi-restricted diet (optimal proportions).” and “In discussion to draw more attention to the possible disadvantages of low-carbohydrate diet in people with diabetes.”

Role of starch versus mono- and disaccharides

Replacement of carbohydrate with fat or, in some cases, with protein, is beneficial in both types of diabetes leading to better glycemic control, weight loss, cardiovascular risk markers, and reduction in medication. This is what we know. That is what is established in well-controlled experiments in individuals with diabetes (points 1, 3, and 10). The evidence does not contain strong data on which carbohydrates should be removed (or even what the effect of different fats of protein might be). On first principle, glucose is of greatest importance in diabetes. The sudden interest in fructose and sucrose as unique types of carbohydrate has made the discussion quite controversial.

Both the scientific [92], [93] and popular literature [94] have been unrestrained in attributing harm to fructose. Generally, fructose is known to have unique effects compared with glucose, although most of these are seen on a high-carbohydrate diet [95] and there may be little difference as carbohydrate is lowered. It is likely that on a low-carbohydrate diet, most fructose that is consumed will be converted to glucose. We have provided a perspective on the metabolism of fructose [96] where we emphasize its integration into general carbohydrate metabolism. The fact that up to 60% of ingested fructose can be converted to glucose makes the analysis of which sugar does what very difficult.

The definitive experiment, testing whether removing fructose is preferable to removing glucose in the implementation of a low-carbohydrate diet has never been performed. This is presumably due to the poor acceptance of low-carbohydrate diets in general [4]. One study showed that glycemic response was lower after ingestion of a low-starch meal with 43% total carbohydrate and high levels of fruit compared with a high-starch, high-carbohydrate meal or a 40% carbohydrate “typical American meal” [97].

There was also, as expected, a lower 24-h integrated serum insulin response. The results demonstrate the value of specifically removing starch, although it was not determined whether removing sugar would be equally effective or better. As above, this group has also shown good results simply by reducing glucose (point 3).

Because of the limited insulin effect, it was once thought that fructose might be an acceptable source of carbohydrate, but this turned out to be questionable and may actually have a deleterious effects if administered intravenously alone. Analysis of the hepatic metabolism shows that the liver expects the two sugars to appear together [96], fructose (e.g., increases glucokinase activity).

The reviewer's original question is framed in terms of “carbohydrate-semi-restricted diet (optimal proportions).” It is unlikely that there is a general answer. As a guide for the patient with diabetes, the prescription of many agencies to “eat to the meter” seems like a good one.

 

Possible disadvantages of low-carbohydrate diet in people with diabetes

To assess the disadvantages of carbohydrate restriction for individuals with diabetes, one has to ask what the standard is and where it came from. The idea that there is an effective diet of known macronutrient composition, one tested in long-term, or even short-term trials, that is beneficial in treating diabetes is implied by the question. To our knowledge, no such diet exists. The more dietary carbohydrate, the more medication will be required (point 11). The disadvantage to a low-carbohydrate diet, as in any intervention, will rest with individual choices. Low-carbohydrate diets generally have better compliance (point 5) but individuals vary in tastes and assessment of risk–benefit perceptions.

The flipside of the benefit from reduced medication (point 11) presents a real potential disadvantage. Because of the effectiveness of carbohydrate restriction on glycemic control, there is a danger of hypoglycemia for those patients on glucose-lowering medication. It is recommended that medication be reduced in advance of initiating a low-carbohydrate diet. Personal communications suggest that there are a variety of strategies for reducing insulin or other drugs. Whether the patient (or the physician) knows this is potentially serious question. Instructions for the study in reference [80], for example, provide the following guide:

“Metformin was continued for the duration of the study unless the participant or his/her doctor requested it be lowered, at which point the dose was cut in half or discontinued completely. Sulfonylurea doses were reduced in half if the entry HbA1c was <7.5% or discontinued if the participant was on a minimum dose. Sulfonylurea was discontinued if predinner glucose levels went below 110 mg/dL despite prior dose reduction; thiazolidinediones were continued for participants with starting with a HbA1 c above 7% and discontinued for those with starting HbA1 c below 7%.”

 

Conclusion and recommendations

What evidence would be required to change the current recommendations for dietary treatment in diabetes? Evidence-based medicine tends to emphasize RCTs as a gold standard. Such absolute requirements, however, are unknown in any scientific discipline. As in a court of law, science admits whatever evidence is relevant [98]. Following the legal analogy, one has to ask: Who decides on the admissibility of the evidence? The parody by Smith and Pell [99] has been described as both funny and profound in illustrating how there is not a single type of experiment that fits every scientific question. Given the current state of research funding and the palpable bias against low-carbohydrate approaches [4], it is unlikely that an RCT can be performed that will satisfy everybody. The seriousness of diabetes suggests that we have enough evidence of different types to reevaluate our current recommendations for treatment.

This review has described 12 points of evidence based on published clinical and experimental studies and the experience of the authors. The points are supported by established principles in biochemistry and physiology and emphasize that the benefits are immediate and documented while the concerns about risk are conjectural and long term.

We would recommend that government or private health agencies hold open hearings on these issues in which researchers in carbohydrate restriction can make their case. We think that traditional features of the analysis of evidence such as vigorous cross-examination should be part of the process. We suggest that open discussion with all sides contributing will be valuable. The seriousness of diabetes suggests that a bench decree will be inappropriate.