The folly of the American Diabetes Association’s dietary recommendations
An eminent diabetologist, Dr Ralph DeFronzo, described the deterioration into diabetes in the Lilley Lecture of 1987, the pre-eminent forum of accumulated medical wisdom and insight:
…
there are two primary defects responsible for the pathogenesis of
non-insulin-dependent diabetes mellitus (NIDDM). In some NIDDM patients the primary defect
starts at the level of the β-cell and manifests itself as an impairment in
insulin secretion; these individuals are represented by the lean diabetic
patient. In other NIDDM patients the
primary defect starts as an impairment in tissue (muscle and liver) sensitivity
to insulin; these individuals are represented by the obese diabetic. However, whichever defect – diminished
insulin secretion or insulin resistance – initiates the development of NIDDM,
it will subsequently lead to the emergence of the second abnormality (DeFronzo
1987).
Dr DeFronzo’s first category includes insulin-dependent diabetics who have no β-cells, and are generally lean and sensitive to insulin, at least initially. But what he describes is startling: whether there is diminished insulin secretion or insulin resistance, the second abnormality will emerge. The insulin-deficient become insulin-resistant and the insulin-resistant become insulin-deficient! Then, in both cases, high blood sugar becomes an agent of further destruction and it’s startling to learn that high blood sugar is almost universal in the diabetic population receiving medical care:
…
there appears to be abundant evidence from both animal and human studies that
sustained chronic hyper-glycemia can lead to the development of defects in both
insulin secretion and insulin action.
Furthermore, it appears that these deleterious effects of hyperglycemia
are mediated by downregulation of the glucose transport system. Recognition of
the important pathogenetic role of hyperglycemia per se in the evolution of NIDDM has important therapeutic
implications.
Over the last decade, we have performed oral glucose (100-g) tolerance tests (OGTT) in many normal-weight NIDDM subjects. All subjects had fasting hyperglycemia (mean ± SE 148 ± 8). After glucose ingestion, the plasma glucose concentrations were >200 mg/dl at all time points …
Yet, inexplicably, only completely out-of-control diabetics are treated so as to achieve normal blood sugars:
Thus, in the poorly controlled diabetic with fasting plasma glucose levels >200 mg/dl it seems reasonable to institute a short course of intensive insulin therapy (4-6 wk) or severe caloric restriction (10-14 days under close supervision by a physician) to restore normoglycemia and reverse the deleterious effects of hyperglycemia on cell function (DeFronzo 1987).
The treatment suggested to all diabetics is the low-fat American Diabetes Association diet, yet this diet causes the hyperglycemia which eventually makes all diabetics both insulin-dependent and insulin resistant! There is no doubt that this is still the case today, in spite of treatment with insulin-stimulating sulfonylurea drugs and the insulin-sensitizing drug metformin. The findings of the UK Prospective Diabetes Study included that:
The study showed an initial increase in β-cell function (from 46 to 78%) at 1 year in subjects on a sulfonylurea, followed by a steady decline in function to 52% at 6 years. Subjects on diet only (n = 486) exhibited a gradual decline in β-cell function of ~ 4% per year. Insulin sensitivity only changed in subjects on metformin (n ~ 159), increasing from 51 to 62% at 1 year and remaining at 62% at 6 years (Wallace 2004).
In other words, the American Diabetes Association diet alone destroys β-cells at the rate of about 4% per year, and the diet plus a sulfonylurea drug improves β-cell performance initially but then destroys them at a rate near 5% per year. At these rates, diet alone will destroy all β-cells in 11 years, while diet plus a sulfonylurea drug will destroy them in 16 years.
Furthermore, physicians are slow to treat high HbA1c levels in the way that Dr DeFronzo suggested way back in 1987. At the 2006 Scientific Sessions of the American Diabetes Association, we learned that:
The
majority of patients with Type 2 diabetes have very high HbA1c levels, and it
takes physicians months before they intensify oral antidiabetic therapy … a
study of 9,416 patients … [showed that] Mean HbA1c was 8.4% at baseline when
therapy was initiated. Only 33% had
levels at or below the
That Dr DeFronzo’s prescription of caloric restriction is effective was demonstrated in a study of obese, hyperglycemic diabetics. “Severe” caloric restriction quickly lowered their fasting blood sugar from 326mg% (!) to 150mg%, and their β-cells showed significant recovery (Stanik 1980). However, a similar study showed that subjects with diabetes of less than two years duration responded far better than subjects who’d had diabetes for more than five years: caloric restriction brought the less-than-two-year diabetics down to 119mg%, but the more-than-five-year group averaged 175mg% (Nagulesparan 1981), for the reason that chronic hyperglycemia had killed more β-cells in the diabetics of long standing.
β-cells are not the only cell type to be damaged by high blood sugar, and the Diabetes Control and Complications Study showed that this damage causes the diabetic complications of blindness, cataracts, heart attack, stroke and the amputation of limbs. An exacting study has described the relationship between high blood sugar and the risk of various complications (Stratton 2000):
The graph at bottom left shows that risk of amputation or death from peripheral vascular disease doubles between an HbA1c of 5.6% (like mine) and the low end of “tight control” at 7%. But recall that “67% [of Type 2 diabetics] had HbA1c levels approaching 10%” (Kerr 2006), and we see that the risk for these people is more than 10 times higher. But anyone who’s prepared to avoid sugars, starches and starchy vegetables can maintain an HbA1c of 5.6% like mine.
So why does the
It’s a mystery.
Ancient studies showed insulin sensitivity increases with increase of carbohydrate in the diet (Himsworth ), but other studies demonstrate that cells remodel themselves to burn the proportion of carbohydrate and fat in the diet, so insulin sensitivity will inevitably be “worse” in cells adapted to burning fat ().
The hoary old knee-jerk fear that fats promote heart disease must play a part. This fear, implicit and unquestioned in the preamble of every low-fat diet study, is, in fact, irrelevant. Diabetic complications rob the diabetic of vitality, the enjoyment of life, and eventually of life itself; it is the “macrovascular complication” of heart disease which kills most diabetics. Hyperglycemia accelerates the development of macrovascular complications, so prescribing a diet which guarantees hyperglycemia in the interests of avoiding heart disease is a protracted oxymoron. This deadly error causes the complications of diabetes which are, then, iatrogenic diseases, caused by the “Medical Nutrition Therapy” prescribed to treat it. The lesson of my experience is crystal clear and unequivocal: avoid carbohydrates, avoid hyperglycemia.
Numerous studies in the medical literature agree that treating insulin-resistant conditions with a low-fat, high-carbohydrate diet worsens features of metabolic syndrome (Gerhard 2004); and the Insulin Resistance Atherosclerosis study established that 92% of Type 2 diabetics are insulin resistant (Haffner 1999). The prestigious American Association of Clinical Endocrinologists (AACE) goes so far as to say: “Of greatest importance is the avoidance of low fat-high carbohydrate diets unless weight loss is also occurring. The more insulin resistant an individual is, the more insulin they must secrete in order to maintain normal glucose homeostasis, as a consequence, in the absence of weight loss, manifestations of the Insulin Resistance Syndrome will be accentuated when insulin resistant persons increase the amount of carbohydrates in their diet (50)” (AACE on IRS).
Even insulin-sensitive people don’t do that well on such diets. Dr Gerald Reaven fed 10 healthy post-menopausal women either 60% or 40% carbohydrate diets. The higher carbohydrate level worsened insulin resistance and elevated unhealthy blood fats: “Because these changes would increase risk of ischemic heart disease in postmenopausal women, it seems reasonable to question the wisdom of recommending that postmenopausal women consume low-fat, high-carbohydrate diets” (Jeppesen 1997). And these were healthy women.
When Dr Reaven made a similar
experiment in Type 2 diabetics, he found the same adverse changes. He compared the 20% fat, 60% carbohydrate
ADA-pattern diet which included 10% of calories as sugar, as the
Although
plasma fasting glucose and insulin concentrations were similar with both diets,
incremental glucose and insulin responses from
In other words, they did better on the lower carbohydrate diet and got worse on the ADA-pattern diet; but both diets provoked “glycosurea”, the spilling of sugar in the urine, which means that both diets elevated blood sugars far beyond the normal range into the territory where kidney deterioration and progression of the other diabetic complications is inevitable.
Similarly, Dr Arbhimanyu Garg performed a similar study in 42 Type 2 diabetics and found that, “In NIDDM patients, high-carbohydrate [55% carbohydrate, 30% fat] diets compared with high-monounsaturated fat [40% carbohydrate, 45% fat] diets cause persistent deterioration of glycemic control, as well as increased plasma triglyceride and very low density lipoprotein levels, which may not be desirable” (Garg 1994).
There are dozens of such studies,
and their results are consistent: higher carbohydrate diets cause deterioration
of diabetic control and elevation of heart risk factors, while lower levels of
carbohydrates improve all aspects of the condition and lower risk factors for
heart disease. Yet the
The
The
The
message today: Eat more whole grains!
Whole grains and starches are good for you because they have very little
fat, saturated fat, or cholesterol. They
are packed with vitamins, minerals, and fiber.
Yes, foods with carbohydrate -- starches, vegetables, fruits, and dairy
products -- will raise your blood glucose more quickly than meats and fats, but
they are the healthiest foods for you.
Your doctor may need to adjust your medications when you eat more
carbohydrates. You may need to increase your activity level or try spacing
carbohydrates throughout the day.
On
average Americans eat around 40-45% of our calories as carbohydrate. This is a moderate amount of carbohydrate,
not high. Currently some controversy about
carbohydrates is raging due to a few new diet books. These books encourage a
low carbohydrate, high protein and moderate fat intake. These diets are not in synch with the
American Diabetes Association nutrition recommendations, which are based on years
of research and clinical experience. In
addition, these trendy diets are hard to follow year after year.
A
way to see how carbohydrates affect your blood glucose is to monitor your blood
1&1/2 to 2 hours after meals. Checking your blood glucose at this point
tells you how high your blood glucose went from the carbohydrates you ate. For
good diabetes control, keep your after-meal blood glucose levels at 180 or
below (ADA
Diabetes Food Pyramid FAQ).
In other words, forget the Zone (40% carbohydrate) and South Beach (45% carbohydrate) diets, go for maximum carbohydrates! But there are troubling questions about the effectiveness of this approach for weight loss: “overweight and obese women assigned to follow the Atkins diet, which had the lowest carbohydrate intake, lost more weight and experienced more favorable overall metabolic effects at 12 months than women assigned to follow the Zone, Ornish, or LEARN [very high carbohydrate] diets” (Gardner 2007). More troubling questions arise over the effects of the high-carbohydrate diets on lipid profiles:
Table. Changes From Baseline at 12 Months*
|
Outcome |
Atkins |
Zone |
LEARN |
Ornish |
|
Weight, kg |
-4.7 |
-1.6 |
-2.6 |
-2.2 |
|
LDL, mg/dL |
0.8 |
0.0 |
0.16 |
-3.8 |
|
HDL, mg/dL |
4.9 |
2.2 |
2.8 |
0.0 |
|
Triglycerides |
-29.3 |
-4.2 |
-14.6 |
-14.9 |
|
Systolic BP |
-7.6 |
-3.3 |
-3.1 |
-1.9 |
*LDL, low-density lipoprotein cholesterol;
HDL, high-density lipoprotein cholesterol; and BP, blood pressure.
Source: JAMA. 2007;297:969-977.
In an interview, the researcher, Dr Christopher Gardner, remarked that "After all those [statistical] adjustments... everything that was still significant favored Atkins, more weight loss than Zone, better triglycerides than Zone, better HDL than Ornish, better blood pressure than all three … there is something really interesting, I think, about carbohydrates and the emphasis really seems to be on cutting back those simple carbs" (Lie 2007).
The science behind the recommendation …
In the ADA’s most recent summary of the research underlying their nutritional recommendations is their Evidence-Based Nutrition Principles and Recommendations for the Treatment and Prevention of Diabetes and Related Complications (Franz 2002), there is not a shred of high-quality research evidence cited in support of their opinion that “Carbohydrate and monounsaturated fat should together provide 60–70% of energy intake”, certainly no randomized controlled clinical trials which are the gold standard among research studies. It is the consensus opinion of the ADA Expert Panel. The only high-quality evidence cited relates to how destructive very high carbohydrate diets are!
In their current update, Nutrition Recommendations and Interventions for Diabetics (ADA 2006), they add their opinion that “Low-carbohydrate diets (restricting total carbohydrate to <130g/day) [about 25% of calories in a 2000 calorie diet] are not recommended in the treatment of overweight/obesity … [or] diabetes”, and cite the Institute of Medicine’s Dietary Reference Intakes: Carbohydrate, Fiber, Fat etc. (IOM DRIs), in which it is made clear that the amount of glucose necessary to power the obligate glucose-consuming tissues like the brain and blood cells is about 130 grams. But this is a specious rationale! By definition, RDAs are for healthy people, being “the average daily dietary nutrient intake level sufficient to meet the nutrient requirement of nearly all (97 to 98%) healthy individuals in a particular life stage and gender group” (p 22). Diabetics are not healthy people. My experience is that carbohydrates are a metabolic poison for me, and, moreover, a poison which steadily worsens the condition, and that 130 grams per day is far too much; my diet contains nearer 30 grams/day, or 6%.
A second difficulty with this
citation is that the Institute of Medicine is at pains to point out that the
absolute requirement for carbohydrates in human nutrition is actually zero, “provided that adequate amounts of
protein and fat are consumed”, but they wish to err on the side of caution
because although various populations thrive on very low carbohydrate diets, “a
comparison with populations ingesting the majority of food energy as
carbohydrate has never been done” (p
275). This isn’t scientific
evidence, it’s an opinion. We think that
you should eat lots of carbohydrates, and the scientific evidence is, er, other
people think the same thing. It’s a
different world, in which there is resort to Aristotelian logic as though
Francis Bacon had never formulated the scientific method. We reason what should be, and then assert
it. And where, as Bacon put it: “For
what a man had rather were true, he more readily believes.” Without a doubt, we have followed
A close reading of the material
intended for medical professionals gives implicit approval to a 25%
carbohydrate diet, but the material intended for the public says: go all-out
for carbohydrates. The material intended
for medical professionals acknowledges the destructive potential of high
carbohydrate diets, but the material intended for the public says no such
thing. Is this because the
The
That this prescription makes high
blood sugars inevitable is implicit in the advice from the ADA FAQ: “For good
diabetes control, keep your after-meal blood glucose levels at 180 or
below.” A glucose level of 180mg% is not
only egregiously high, it’s the renal threshold where glucose begins to appear
in the urine. But even this admonition
is honored mainly in the breach. A
typical study gave the mean HbA1c of 17 diabetics in the
experimental groups at 9.9%, and that of the control group at 10% (Georgopoulos
1995); an HbA1c of 10% corresponds to a blood sugar of
247mg%, which guarantees chronic glycosurea.
In Dr Gerald Reaven’s diet comparison excerpted above, “24-hour urine
glucose excretion was significantly greater (55 versus 26g/24 hours p<0.02)
in response to the low-fat, high-carbohydrate [
This is where the mystery deepens:
the clear and present danger is diabetic complications caused by elevated blood
sugars (as graphically illustrated above), yet the ADA dietary recommendations
cause such high blood sugars that sugar appears in the urine, even when
monounsaturated fats replace some carbohydrate.
As well as making complications inevitable, Dr DeFronzo has pointed out
that hyperglycemia causes both increased insulin resistance and the progressive
loss of β-cell function which eventually renders insulin necessary. In other words, the hyperglycemia attendant
upon the
Yet the low-carbohydrate diet I eat
which renders my blood sugar normal is “not in synch” with the
The
A claim that low-carbohydrate diets
are ineffective for glycemic control would be mendacious, but the
|
Effect of Carbohydrate Intake on HbA1c in Seven Low-Carbohydrate Diet
Studies
|
||||||||||||
|
|
Number |
Type |
Duration
wks |
BMI
pre |
BMI
post |
∆BMI%/wk |
Carbs
pre |
Carbs
post |
HbA1c
pre |
HbA1c
post |
HbA1c
final |
|
|
10 |
2 |
2 |
39.8 |
39.0 |
1.0% |
40% |
4% |
7.3 |
6.3 |
5.6 |
projected |
|
|
8 |
2 |
5 |
31.0 |
30.4 |
0.4% |
55% |
20% |
9.8 |
7.6 |
7.6 |
|
|
|
9 |
2 |
6 |
36.3 |
33.5 |
1.3% |
55% |
10% |
9.6 |
6.8 |
6.8 |
|
|
|
19 |
2,
no OHAs* |
8 |
27.9 |
26.5 |
0.6% |
55% |
25% |
9.9 |
8.1 |
8.1 |
|
|
|
9 |
2,
OHAs* |
8 |
29.6 |
29.3 |
0.1% |
55% |
25% |
9.2 |
7.8 |
7.8 |
|
|
|
30 |
20x2,
10x1 |
86 |
28.7 |
27.9 |
0.0% |
55% |
6% |
7.9 |
5.7 |
5.7 |
|
|
|
16 |
2 |
24 |
36.1 |
32.0 |
0.5% |
55% |
20% |
8 |
6.6 |
6.6 |
|
|
|
21 |
2 |
16 |
42.2 |
39.4 |
0.4% |
40% |
10% |
7.5 |
6.3 |
6.3 |
|
|
|
*OHAs:
prior treatment with Oral Hypoglycemic Agents
|
||||||||||||
Left:
Interestingly, these seven low-carbohydrate (<25% of calories) studies show
that there is a strong correlation between the percentage of calories as
carbohydrate and HbA1c at the end of the study; “R2 =
0.8374” means that, statistically speaking, the differences in the percentage
of carbohydrate calories accounts for about 84% of the change in the values of
HbA1c. The rest of the variation likely
comes from duration of diabetes and the degree of obesity. Right: Weight loss no doubt contributes to
the lowering of HbA1c for weight loss occurred in all 7 studies, but
the absolute intake of carbohydrate was far more important. A review of very low-calorie diet therapy for
obese diabetics points out that “Metabolic benefits occur quickly with only
modest weight reduction, suggesting that calorie restriction plays a more
critical role [than the absolute amount of weight lost]” (Henry
1991). It seems that, in
low-carbohydrate diets, caloric restriction is a necessary but not sufficient
condition for improvement of the diabetic state, and that it is the degree of restriction
of carbohydrate which predicts the degree of benefit which will accrue: only the diets with less than 8% of
carbohydrate achieved an HbA1c of less than 6%.
This is in good agreement with my experience. My diet contains about 6% carbohydrate, like Dr Richard Bernstein’s diet upon which it is based, and my HbA1c is 5.6%. Before Dr Bernstein published his book in 1998, I ate about 25% of calories as carbohydrate and my HbA1c wandered between 7 and 8%:
A graph of my HbA1c scores before and after I adopted a low-carbohydrate diet
Conversely, high-carbohydrate diets improve HbA1c only if they are calorie-restricted, and this effect is barely strong enough to be clinically useful for Type 2 diabetics. For example, when Dr Leonie Heilbronn restricted calories by 30% in two diets containing 60% carbohydrate which were made up of either of low or high Glycemic Index foods and fed them to two groups of diabetics for 8 weeks, she found that “Urinary glucose excretion was not significantly changed … in either group” (Heilbronn 1999). There is a trivial improvement in the low Glycemic Index group, but any urinary glucose excretion means microalbinurea (the loss of protein through the kidneys) is not far behind, which will be followed by macroalbinurea and eventually by kidney failure. The literature is rife with such studies (Gumbiner 1998, Parker 2002,), performed by researchers who are cavalier to the point of indifference. Where are the Ethics Committees? How can a literature search be so bungled that low-carbohydrate, blood sugar-normalizing diet studies are overlooked?
And sometimes, we just don’t see the forest for the trees. Dr Cecilia Low compared calorie-restricted diets with 10% and 70% carbohydrate content and found clear advantages for the low-carbohydrate diet over the 6-week study period in obese Type 2 diabetics, yet she dismissed the entire approach with: “it is impractical to consume weight-maintaining diets composed of 70% fat” (Low 1996). Never mind that the low-carbohydrate group had just completed four weeks of “re-feeding” doing precisely that, albeit on a liquid-formula diet! Never mind that there is a whole population of free-living diabetics who approach this level routinely, and consequently have normal blood sugars.
On the
The graphs show that the rate of
progression of retinopathy is lowest at an HbA1c of 5.5%, but the
rate of severe hypoglycemia is highest, about one episode per year. Severe hypoglycemia is very risky because you
may make catastrophic errors of judgment, lose control of your car, have a
seizure or worse. However, diabetic
complications consequent on high blood sugars cause blindness, limb amputation,
kidney failure and accelerated heart disease.
Insulin-dependent diabetics constitute only 10% of the diabetic
population, yet the argument that tight control is too risky because of the
attendant hypoglycemic episodes is somehow generalized to all diabetics,
implicit in the
In
the 20 plus years that I have been in practice, only five of my patients have
had severe hypoglycemia causing loss of consciousness. Two of these people were
eating excessive amounts of carbohydrate and three made major mistakes such as
taking the wrong type of insulin. I’m
sure this is a far cry from the incidence of severe hypoglycemia among patients
of high carbohydrate practitioners (Bernstein
Interview).
Severe hypoglycemia is vanishingly rare on low-carbohydrate diets. So what’s the secret? For one thing, there’s ketosis, in which two- to four-carbon fragments of fats are formed in the liver when the diet is low in carbohydrates – these ketones can directly nourish many of the tissues, including some brain structures which would otherwise burn only glucose, so that when the blood sugar drops too low, the brain runs mostly on ketones (Johnson 1978). So a low-carbohydrate diet protects against hypoglycemia. The widespread prejudice against ketones and ketosis stems from the widespread misunderstanding that ketones and ketosis only occur in diabetic ketoacidosis, which is a deadly consequence of too little insulin in insulin-dependent diabetes.
Interestingly, it is the ADA’s own
much-vaunted research, used so much to justify their low-fat diet, which has
demonstrated that a low-carbohydrate diet lowered HbA1c to normal in
8 overweight Type 2 diabetics. The study
was funded by the
Left: Mean plasma glucose concentration before (▲) and after (●