Dietary recommendations in diabetes mellitus (DM) have changed over the years. Prior to 1980, they avoided sugar, and were enriched in fat. However, due to concerns raised over hypercholesterolemia, and its relationship to dietary fat, a shift occurred in the 1980s to diets that included more complex carbohydrates (starches). It is ironic that during the same period of time, awareness of the varied glycemic effects of foods occurred. The fact that different starches caused varied increases in blood sugar when they were eaten has been very gradually assimilated into medical practice. It has only been since the mid-1990s that clinical studies evaluating the utility of a glycemic index in ranking dietary starches was evaluated.

We can stratify starches by glycemic index (GI), but was it important? Epidemiologic work conducted by Walter Willet and his colleagues at the Harvard School of Public Health, indicated that high glycemic diets increased the risk of type II DM (adult onset). The initial study was performed in women (JAMA 277(6):472-7, 1997 Feb12), with a similar study subsequently reported in men (Diabetes Care 20(4):545-50, 1997 Apr). Both showed that in people on a high GI diet compared to those with a low GI diet, those with a high GI diet had a 50% greater risk of developing diabetes.

Though the risk was increased, it was likely that the high glycemic diet was causing a metabolic abnormaility, insulin resistance, to be expressed as clinical diabetes. Thus, attempts to identify and modify this underlying metabolic defect were likely to be important in trying to affect the risk of developing DM. A 1997 report in Science (October 17; 278: 419-424) indicated that 40% of Americans would develop IR or DM by the age of 65, and 50% would by the age of 80. This indicated that the genetic predisposition to IR and DM was indeed high in our genetically heterogenous population.

The surest way to get fasting insulin levels down (a good clinical tool for assessing insulin resistance), is to use a protein sparing modified fast (PSMF). This has been popularized by the Atkin's Diet (his web site is such a marketing mess, I didn't include it here). Suffice it to say, the main focus is on high protein foods (flesh foods). The major ingredient of a PSMF is to consume 1.5g of high quality protein per kg of ideal body weight. Generally, an ounce of cooked meat (beef, chicken, fish, pork) will provide 7g of protein. Thus, most people will be consuming 12-16 ounces of "meat" a day. Eggs count as one ounce of protein. Atkin's permits cheese, but most PSMF don't permit this. The major metabolic effect of such a diet is to limit carbohydrate (CHO) to less than 40 g a day, so one stays in ketosis. This allows stable energy homeostasis without requiring insulin to maintain blood sugar levels. As you can see, cheese isn't a major problem, as it has little CHO (but alot of fat. However, fatty acids are chewed up without a problem when fatty acid oxidation is heightened during ketosis). Most of the CHO on a PSMF comes in the form of vegetable servings. Generally 2-3 servings a day are permitted (1/2 a cup of cooked veggies or 1 cup of raw is the basic serving size which contains 25 kcal in the form of 5 g of CHO). Thus, the veggies keep the CHO in the desired range so fat burning, i.e. ketosis, is maintained.

The problems with PSMF include: they can lead to lean tissue losses, especially in people who are less than 40 pounds overweight, they haven't been studied for periods of longer than 16 weeks, they are quite restricted in terms of food choices, and can have significant side effects (e.g. 25% of people on a PSMF will develop gall stones in the first 8 weeks of the diet. This excludes people with pre-existing stones). Though PSMF work, their long term success is not impressive (at best, 10% will maintain significant weight loss at 5 yrs after the diet). Also, they require great motivation to adhere to even in the short term, let alone over a prolonged period of time. Finally, using such a diet in people who are near IBW hasn't been well studied, and is likely to be associated with significant lean tissue losses (which has always correlated with the worst of possible PSMF complications, sudden cardiac death).

A less drastic approach, that can be sustained over the long run is to get rid of the starches, but permit all the rest. Though fruits and vegetable have CHO, they are in limited quantities, and as we have learned from the GI, they don't stress the insulin system like starches. After all, this is the critical issue. How do you get a person with a defective glucose disposal system (i.e. insulin resistance) to maintain a normal blood sugar? Since starches are the major foods with high GI, keeping them out of the diet, keeps the foods in ones diet within their means of glucose disposal. Meats, dairy, legumes and nuts do not have marked effects on blood sugar, just as fruits and vegetables. Thus, these foods are the essence of a no CHO diet.

A recent review of medline (2001) using glycemic index as a search term shows several clinical studies. Most of those which have created a sufficient difference between high and low GI diets, have shown impact on glycohemoglobins (a measure of diabetes blood sugar control). When a minimal change has been made (e.g. a trial where the only change was at breakfast where a low GI cereal was exchange for a high GI cereal), the effect on diabetes control hasn't been apparent. My clinical experience has taught me, that a low GI diet is hard to effectively implement due to its complexity. The fact that even what are referred to as "whole grain" in this country have been processed to the point that the starches are rapidly absorbed, complicate the picture. In addition, real German Pumpernickel (a truly low GI bread), not only has a markedly different taste than any bread you have ever eaten, it also is quite expensive. When you just eliminate the starches en total, the goals of the patient are much clearer.

Thus, no rice, no potatoes, no pasta, no bread, no crackers, no cookies, no snack foods, no pancakes nor waffles isn't hard to understand. In addition, they can eat all the meats, fruits, vegetables, legumes (the only truly low GI starch for unclear reasons), dairy (except cheese, which I usually refer to as "poison"), eggs and nuts, give a sufficient basis for most to live on. It is sometimes quite striking that when you reveiw such a diet with your patients, you will find that they live on the starches, and actually eat few of the low GI foods. In such people, even partial compliance will permit them to get off insulin or cut down on medications. As always, you can't ignore exercise, but since it does have a limited effect on insulin resistance (see NEJM...), you have to take the stress off the glucose disposal system. After all, humans have only been eating starches for the past 10,000 years, when agriculture was first started in the "Golden Cresent" (current day eastern Turkey and northern Iraq). The genes which we carry putting us at risk for insulin resistance, likely came with us when our ancestors left Africa (1 million years ago if you believe the archeologic evidence, 100,000 years ago if you rely on mitochondrial DNA analysis). Getting back to our roots, may be the best medicine we can offer people!

A GeneralNutrition Link