Tuesday, April 7, 2009

Fat and Diabetes

The current dogma among many people, including many medical people, is that the current "diabetes epidemic" is caused by high-fat diets, which cause obesity, which causes diabetes. Thus, if someone is overweight, that person is urged to go on a low-fat diet to lose weight and not progress to type 2 diabetes.

Similarly, a person already diagnosed with type 2 diabetes is often put on a low-fat, high-carbohydrate diet in an effort to help that person lose weight. Then when their blood glucose (BG) levels go high, they're given medication to bring the BG levels down.

The rationale for this approach is based on decades-old studies showing a relation between high-fat diets and heart disease. Because people with diabetes usually die from heart disease, it was thought they should eat less fat. Eating less fat usually means eating more carbohydrate.

Now, people are beginning to criticize the science relating the high-fat diets to heart disease (some people never believed it). And in any case, a relation between factor X and factor Y doesn't always imply causation. Drink a lot of beer and you'll get drunk. You'll also pee a lot. Hence peeing a lot is related to being drunk. But peeing a lot doesn't cause intoxication. A third factor, the alcohol, caused both the intoxication and the increased urination.

Nevertheless, the popular perception remains: eat fat and you'll get fat and get diabetes.

Hence I was interested when I recently came across a study showing that mice who were predisposed to diabetes were protected from getting diabetes when they ate a high-fat carbohydrate-free diet.

Research in the 1980s had shown that substituting protein for carbohydrate protected db/db mice from getting diabetes.
The db stands for diabetes, because these mice are predisposed to becoming obese and developing diabetes.

It has been shown that they have defective receptors for the hormone leptin, which is one hormone that controls appetite. Because the leptin can't work properly to turn their appetites down, they have voracious appetites and become obese.

And the effect of substituting fat for carbohydrate had previously been demonstrated in another strain, the New Zealand Obese (NZO) strain, which is considered a model for metabolic syndrome and type 2 diabetes. These mice show insulin resistance, high triglyceride levels, high blood pressure, and a low first-phase insulin response. When they get over a certain weight, their beta cells begin to die.

The new study showed the same preventive effect of fat on the db/db strain of mice.

In the cited study, performed by a German group, mice were divided into three groups: normal mouse diet (5.1% fat, 58.3% carbohydrate, 17.6% protein), high-fat diet (14.6% fat, 46.7% carbohydrate, 17.1% protein), and carbohydrate-free high-fat diet (30.2% fat, 0% carbohydrate, 26.4% protein). The mice had free access to the food and water.

The mice on both the high-fat diets did gain weight faster than the mice on the control diet. A strain of lean mice on the same high-fat diets gained more weight than control mice, but not enough to be considered obese.

The interesting thing was that the mice getting extra fat plus carbohydrate not only gained more weight than the control mice, but they also became diabetic faster. However, the mice getting a lot of extra fat but no carbohydrate got obese (fatter than the mice getting a lot of fat plus carbohydrate), but their BG levels were much lower.

They were also producing as much insulin at the end of the study as they were at the beginning, an indication that their beta cells were still healthy. Histological studies of the beta cells showed that the mice on the carbohydrate-free high-fat diet had larger, healthier beta cells than the mice in the other two groups.

Of course we all know the limitations of mouse studies. Something that works in mice doesn't always pan out to work well in humans. Nevertheless, this interesting study shows that a carbohydrate-free high-fat diet can greatly reduce the tendency to become diabetic in mice that have a strong genetic predisposition to do so.

It also shows that mice without a tendency to get diabetes can gain weight on a high-fat diet without getting diabetes.

The diet that was higher in fat than normal but still contained almost 47% carbohydrate diet did what people tell us a "high fat" diet will do. It accelerated the rate of both obesity and diabetes in a highly susceptible population.

Some researchers call a 47% carbohydrate diet a "low carb" diet because it's lower than the 60% carbohydrate diet that has been recommended by groups such as the American Diabetes Association. But it wasn't low enough for the db/db mice.

When carbohydrates were totally eliminated, the effect on the diabetes was reversed. Although the mice got obese, they didn't develop as much diabetes.

This is an example of correlation vs cause. The high-fat diet caused obesity, but the obesity didn't cause the diabetes. It was carbohydrate that was causing the diabetes, probably by killing beta cells that had a genetic tendency to be stressed.

What this means for humans is not clear. Not only do humans not always react the same as mice, but it's unlikely that anyone would want to remain on a totally carbohydrate-free diet for a long time. Even the Atkins induction diet, which has the least amount of carbohydrate for a month or so, includes a few carbohydrate foods like lettuce.

Many people are able to maintain health on a low-carb diet that includes a variety of low-carb vegetables such as broccoli, cauliflower, leafy greens, and green peppers, and even a few fruits such as berries. Going totally carbohydrate free would be difficult indeed.

However, this study suggests that it's not fat that is real culprit in causing diabetes. It's carbohydrate. The worst scenario seems to be when you mix a lot of fat with a lot of carbohydrate. This is exactly what the current "standard American diet" does. French fries (carbohydrate) cooked in oil (fat). Hamburgers (fat) in large buns with soda (carbohydrate).

If people are like mice, then adding even more fat to a diet that still contains at least 46% carbohydrate would indeed accelerate the rate of weight gain and diabetes in people with a genetic susceptibility to diabetes. People without the genetic tendency would gain a little weight, but probably not a lot, and they wouldn't get diabetes.

But how about a much lower carbohydrate diet? The low-carb diet supported by Dr. Richard Bernstein, recommends 30 g of carbohydrate a day. Others recommend 50 or up to 100 g. The percentages would depend on how many calories you're eating. On a 2000-calorie diet, 30 g would be 6%, 50 g would be 10%, and 100 g would be 20%, all far below the 46% the mice with accelerated diabetes rates were getting.

Which carbohydrate level would give the results shown in the mice when they went on a no-carb diet? No one knows. In the earlier study in db/db mice, although an 8% carbohydrate diet slowed the rate of diabetes, eventually the mice all got diabetes anyway. "Only the carbohydrate-free diet provided effective, long-term therapy," the authors wrote.

And how about the obesity? In the recent study, mice on the no-carb high-fat diet did gain a lot of weight. But the mice were allowed free access to food all day. They probably weren't concerned about low self-esteem from being fat. Humans could be taught about the potential weight-gaining effects of such a diet and counseled to eat only as much as they needed to keep hunger away.

One advantage of fat is that it does suppress appetite, so eating fewer calories on a high-fat diet is often easier than eating fewer calories on a low-fat diet.

One thing this study illustrates is that dogma can change. The science that everyone accepted as true in the past may be proven to be wrong in the future. We need to keep open minds.

Is the dietary advice being given to overweight people who want to avoid getting type 2 diabetes what is actually causing the diabetes epidemic? With counseling and a change in the official dogma about fat and diabetes, could we reverse the "diabetes epidemic"? We don't know. But we can always hope.


I neglected to say above that of course the idea that excessive carbohydrate intake may contribute to diabetes and make existing diabetes control more difficult is not new. Dr. Richard Bernstein has been urging low-carb diets for diabetes for many years, and Gary Taubes wrote a detailed and well-documented book (Good Calories, Bad Calories) supporting the idea that fat is not the enemy, carbohydrates are. Many other people have also supported the idea of low-carb diets for health (for example, Michael Eades in Protein Power).

What was interesting about the mouse study I discussed was the fact that the fat content of the diet was extremely high for mice because of the zero carbohydrate intake, and yet the diet didn't have any of the deleterious effects on diabete onset that the low-fatters would predict. Instead, it was beneficial in a highly susceptible population.


  1. Halleluiah!

    Yes just like most human studies, most mouse studies start with standard mouse chow, which is high carb, and tinker slightly with the fat content. Only when you start to get the carbs down to non-toxic levels do you see the fat metabolism change significantly.

    No doubt you already saw this, but for the benefit of your readers


    THIS is what the last fifty years of dietary advice has been built on.

    Meanwhile Stephan


    has dug up some research showing that the replacement of those nasty saturated fats not only with all those healthy carbs but with Omega 6 oils and fructose may also be making the problem worse. Eat Your Dietician! (grins)

  2. You say
    In the cited study, performed by a German group, mice were divided into three groups: normal mouse diet (5.1% fat, 58.3% carbohydrate, 17.6% protein), high-fat diet (14.6% fat, 46.7% carbohydrate, 17.1% protein), and carbohydrate-free high-fat diet (30.2% fat, 0% carbohydrate, 26.4% protein). The mice had free access to the food and water.

    Er, my maths makes these numbers add up to 81 percent, 78.4 percent, and 56.6 percent respectively.. And the rest of the diet~?

  3. Good question, Helen. The percentages were w/w, or weight/weight, so I think the missing percentages would be water or indigestible fiber.