Diets and BG Levels

Patient Expert

Many of us find that low-carbohydrate (LC) diets work the best to control our blood glucose (BG) levels. Many of us also find that LC diets control our hunger as well, which is very important when you want to lose weight.

And although you can lose weight on various types of diets if you control your calorie intake, controlling your calorie intake can be difficult when you're ravenously hungry most of the day. That happened to me. I lost weight on a low-calorie, low-fat diet when I was diagnosed, but I was ravenous the whole day and finally decided I couldn't live with this for the rest of my life and gradually switched to a LC diet, and the hunger disappeared.

Some people don't feel ravenous on a low-fat diet, and for them any type of diet works for weight loss or simply for health maintenance. But there's evidence that people with insulin resistance, which means most of us as well as people with metabolic syndrome, often a precursor to type 2 diabetes, do better on LC diets. In fact, one paper suggests that doing better on a LC diet might be considered a definition of metabolic syndrome.

I think the key here is that when you have insulin resistance, your BG levels can go very high after eating carbs, and then they come down fairly fast until you have overt diabetes. And I think this rapidly falling BG level triggers hunger. It's logical. If your BG is falling fast, the body wants to correct it before you go low instead of waiting until you're low, when it would take 15 or 30 minutes to correct the low, and by that time you could already be alarmingly low.

When I did mini oral glucose tolerance tests (OGTTs), I found I got ravenous just after the peak BG level, as illustrated in Figure 6 in my book The First Year: Type 2 Diabetes. If you can avoid those peaks and rapid falls, your hunger should abate.

_  _

People without insulin resistance don't see this effect, because their BG levels don't usually go as high after eating.

Thus, for people with insulin resistance, LC diets are beneficial. However, one thing many people on LC diets don't realize is that when you're on a LC diet, you become more sensitive to carbs. This means that an amount of carbs that wouldn't make your BG levels go up much if you weren't on a LC diet can make them go up a lot if you're not.

The most extreme example of a LC diet is starvation, zero carbs as well as zero anything else. People who are starving will test diabetic on OGTTs even if they're not. So that's why when you take an OGTT, for example, if you're pregnant, they tell you to eat a lot of carbs for 2 or 3 days before the test.

Why is this true? One reason is that when you don't eat carbs, your body stops producing the enzymes that metabolize carbs so they can remain in your bloodstream longer. Another reason is that when glucose is in short supply, the body tries to conserve it to use in the brain, which does need some glucose. One way to do this is to increase insulin resistance in the muscles, which can get their energy from fat (and often prefer to).

But is this insulin resistance bad?

I don't think it is as long as you stick to the LC diet. In other words, if you decide a LC diet will work for you, don't eat doughnuts every Sunday or cake when you're invited out, because when you do, your BG levels will most likely go very high.

A paper published in 1989 illustrated some of this.   Unfortunately, although the paper is finally available online, the figures are not. But I tried to read the values from small graphs in the paper, which Kerin O'Dea, the first author, kindly sent me from Australia a few years ago, and put them into a graphing program.

What these researchers did was to put people with type 2 diabetes on 4 diets: high-carb high fiber, high-carb low fiber, LC high fat, and LC high protein. I'll ignore the high-carb low fiber as I think everyone now agrees that this isn't a good approach.

The high-fiber diet (diet 1) "consisted of unrefined cereals, such as wholemeal bread, pasta, brown rice, at least one serving of legumes per day, fruit and vegetables, non-fat dairy products, and very lean meat and fish." It was very low in fat (about 9% total fat).

The LC high-fat diet (diet 3) "was achieved by the use of full-cream dairy products, eggs, fatty meats, butter, margarine, and oil. Leafy vegetables were eaten in usual amounts, but cereals and fruits were restricted." It contained 20% carbohydrate, which would be about 65 to 77 g of carbohydrate (depending on whether or not you subtract fiber) on a 1500-calorie diet. So it was probably not ketogenic.

The LC high-protein diet (diet 4) was low in fat and carbohydrate and very low in fiber. The protein content of this diet was exceptionally high in order to keep the energy content equivalent to that in the other three diets. The calories were "derived predominantly from lean meat (beef, fish, chicken, and pork). Skim milk products were restricted to a combined total of 600 gm/day Leafy vegetables were eaten freely but starchy vegetables, cereals, and fruits were restricted."

Participants weighed and recorded what they ate, and much of the food was provided to them. The researchers did an OGTT on the participants and then repeated it after 14 days on each of the diets. They also measured BG when the participants were eating the test diets both before and after the 14 days. This is a step that is not often done. Researchers usually feed some diet and then measure some parameters at the end of, rather than during, the test period.

The first figure shows the response before and after 14 days of the high-fiber diet.   (Note: the results are presented as mmol/L. To convert to mg/dL, multiply by 18.) Clearly the subjects did better on an OGTT after the high-fiber diet. Figure 2 shows the results from the subjects on the LC high-fat diet. Instead of doing better, they did worse on an OGTT, as if their insulin resistance had increased. The subjects on the LC high-protein diet, like those on the high-fiber diet, did better on an OGTT after the diet (not shown here). It looks as if the decrease was not quite as much as with the high-fiber diet, but they started at a lower level (peaking at about 15 instead of 19). The peaks after the diets were about 15 on the high-fiber diet and about 14 on the high-protein diet.

(Note: In all cases, the Y axis, the vertical one, shows BG levels in mmol/L. I forgot to change the labels on the graphing program, and as there were technical problems involved in getting the graphs to display here, I didn't go back and start all over again.)




Now comes what I find to be the interesting part. Figure 3 shows the BG levels of the participants while they were eating the test diets, not on an OGTT. The participants in the high-fiber diet went up to about 15 mmol/L (270 mg/dL) after eating the high-carb, high-fiber diet before the 14-days on the diet and up to about 13.5 mmol/L (243 mg/dL) after 14 days on the high-fiber diet, when their response to the OGTT was better.



In other words, to get the apparent improvement in insulin resistance, the participants had BG levels over 200 mg/dL after every meal, even ater 14 days on the diet (BGs after lunch and dinner weren't as high as after breakfast, but they were still over 200).

But what about the LC high-protein diet? What were BG levels like when the participants were eating this diet? The results can be seen in Figure 4.



On this diet, BG levels didn't go over about 9 mmol/L (162 mg/dL) before the 14 days on the diet, and were below 7 mmol/L (126 mg/dl) after the 14 days. This would be considered a normal BG reading by most, and you can see the absense of peaks and valleys.   And this occurred despite an extremely high protein intake (for example, about 12 oz of meat for breakfast), which some participants found they couldn't eat.

On the LC high-fat diet, BG levels went up more than they did on the LC high-protein diet, which I found strange because protein can be converted to glucose by gluconeogenesis, and my BG levels will increase if I eat a lot of protein. However, they went up less than on the high-fiber diet (peaking at about 11 mmol/L [198 mg/dL]   instead of 13.5 mmol/L [243 mg/dL] ). But participants on the "LC" high-fat diet were eating more carbs (20 to 30 g per meal and 20% carbohydrate overall, which would be about 75 g per day on a 1500-calorie diet) than most of us on LC ketogenic diets.

Eating lots of fat along with carbohydrate can indeed cause BGs to rise and stay high. I think a diet with less protein than the high-protein diet (which was 68% protein overall) and less carbohydrate than the high-fat diet makes sense. Instead of eating wholemeal bread and mashed potatoes, as the high-fatters did, it makes sense to eat a little more protein and more LC vegetables, which provide fiber as well as vitamins.

The important thing is to test yourself to see which diet works best for you. Figure 5 shows the results of one person who has a BMI of about 20, is physically active, and does not have diabetes.   She was on a LC diet because she felt it was healthy. Her fasting BG levels were close to 100 mg/dL (a sign of prediabetes) and when she did an OGTT using 75 g of glucose and a home meter, she found that her BG levels went up almost to 200 mg/dL. The red line and the dark blue squares show the OGTT on two separate occasions 6 months apart, when she tested using fingersticks. The orange line shows the results on the same day as the red line, but using a continuous glucose monitor (CGM). The CGMs measure interstitial fluid, which lags behind capillary blood by about 15 to 20 minutes, which explains why the curve is shifted to the right.

So then, after having been on a LC diet for years, she decided to try a very low fat, low-glycemic-index vegan diet. The green line shows what her BG levels were on this diet: She never went much above 100 mg/dL, despite eating a low-fat meal that included pearled barley and lentils and had a total of 125 g of carbs, and most of her readings were far below that.

But what if she did an OGTT? She did, with a whopping 75 g of glucose, and the purple line shows those results. Even with this huge bolus of glucose, her BG didn't go much above 100! This is an example of the fact that not eating carbs can make your BG levels look diabetic even if you're not, but eating a high-carb diet can improve your response to an OGTT.



So should we all be on vegan diets? I don't think so. Someone without diabetes, or a diabetic tendency, can eat lots of carbs without increasing BG levels, as shown in Figure 5.   But as shown in the Australian research (conducted on people with type 2 diabetes), if you're diabetic, eating enough carbs to improve your glucose response can make your BGs go extremely high after meals, high enough to cause diabetic complications.

If you've just been diagnosed and you're still producing a lot of insulin, you might be able to tolerate a vegan diet, or just a high-carb, low-fat diet, and if those food choices appeal to you more than those on a LC diet, so you think you could stick to that diet long enough to lose some weight (if you need to lose weight), you could improve your health.

In today's world, we don't have to take anyone's word for it.   We have the tools to keep track of our own health. So use your meter to test foods after meals, and even do an OGTT with glucose (I used to use just 12 g of glucose, and that was enough to make me go up close to 200 mg/dL).

Just remember that the OGTT requires a 3-day carb-up period if you want a true result. If you just want to see how high carbs send you while you're on your LC diet, then of course you don't do a carb-up preparation. The results might convince you that as long as you're on a LC diet, you should avoid carbs as much as you can.