Fuzzy Fats

Fats have been in the news lately, with publication of a meta-analysis mentioned in a previous blogpost showing that there's no statistical evidence that saturated fat is associated with deaths from cardiovascular disease (CVD).

Well, let me qualify that statement. Fats have been in the low-carb news lately. I haven't seen much discussion of this study by people like Dean Ornish who support very low fat diets.

Note that the study showed that they found no evidence that saturated fat intake was associated with CVD mortality. They didn't study whether or not eating other fats were associated with CVD mortality.

But what does saturated fat intake mean? Does it mean the amount of saturated fat you eat? Or does it mean what percentage of your diet consists of saturated fat? These two things can be quite different. Yet many research reports, written by scientists who should be precise about such things, don't make it clear.

Fat intake is often reported as a percentage. For example, low fatters want us to keep our dietary fat under 30% and saturated fat under 10%. But I don't know anyone who goes into a restaurant and sits down with a calculator and a scale to make sure they don't eat more than 30% fat at that meal. (Actually, people with type 1 diabetes used to have to do just that.) Trying to calculate total fat and also saturated fat is even more difficult.

I once bought a little hand-held gizmo that allowed you to punch in your menu and it would, indeed, calculate the macronutrients in the meal. If you ate the same meals over and over again, it might have been useful. But if you ate the same meals over and over again it would have been equally useful to do the calculations on a real computer or even by hand with the aid of a book and then write them down to refer to when you had that meal.

When you ate something different at each meal, you had to locate that food among thousands of foods, decide what description best fit it and how much it weighed, and then input that.

I don't eat the same meal over and over again. I think variety is the spice of eating as well as the spice of life. When it's mealtime, I go to the fridge to see what's there and then try to make something interesting from it.

And the biggest problem I had with any nutritional gizmo was trying to decide which of the myriad choices to input for various foods. Let's say I roast a leg of lamb.

Do I want "Lamb, Australian, imported, fresh, leg, center slice, bone in, separable lean and fat, trimmed to 1/8 inch fat, raw" or do I want "Lamb, domestic, leg, shank half, separable lean and fat, trimmed to 1/4 inch fat, Choice, cooked, roasted, USDA"? I counted 44 different versions of leg of lamb in the nutrition program I have on my computer: Computer Planned Nutrition.

And how do I know how the nutritional information from lambs raised by my neighbors are closer to those of the Australian, New Zealand, or domestic lamb in this program? The lamb I eat comes from free-range animals. Do the animals used in the nutritional program? I don't know.

So even assuming I have time to comb through all 44 choices every time I sit down to eat a slice of lamb (which would probably be cold by the time I figured out the best one), I have no confidence that the lamb I'm eating has the same nutritional composition as the lamb in the program.

Some lambs are fatter than others. Of course I could trim them all evenly. But I could trim the lamb to 1/4 inch or 1/8 inch and then eat or not eat the remaining fat. Or I could leave some of the lamb and a lot of the fat that oozed out of the lamb on my plate.

I see some of the precise calculations some people do with nutritional programs as GIGO: garbage in, garbage out. The computational ability of the computer programs exceeds the accuracy of the data you put in.

The same is true of other foods. Foods have different nutritional compositions depending not only on the particular variety and size but on the condition of the soil, fertilizer, growing season, and so forth.

Furthermore, most nutritional studies rely on people's recollection of what they ate last week or last month or last year. I often can't remember what I had for breakfast, much less last week.

But I digress. I was talking about fat.

Most people don't estimate the percentage of fat in every meal they eat. You can get an estimate of what you need every day by calculating the number of calories you need every day to keep your weight stable, or to lose weight if that's what you're trying to do. Then you can calculate how many calories or grams of a food you should eat each day to reach that percentage.

For example, let's say you're trying to eat about 2000 calories a day with 30% fat. That would be 600 calories of fat. Because there are about 9 calories per gram of fat, that would be 66 grams of fat. Divide that by 28.35 (1 ounce = 28.35 grams), and you get about 2.35 ounces of fat per day, or 1.175 tablespoons (1 oz = 2 Tb). That includes the fat in your meat as well as the oil you cook in or pour on your salad.

But how do you really know how much fat is in the meat you eat?

What if you're eating a lot more or a lot less than 2000 calories a day? What if you're a large man and you're very athletic and you eat 5000 calories a day. Then you could eat 1000 calories of fat (111 grams) of day and still say you were eating a very low fat diet, only 20%.

What if you're a small woman trying to lose weight? You might eat only 1000 calories a day. Then the same amount of fat (1000 calories, or 111 grams) would constitute 100% of your diet, obviously not a likely choice for anyone. If you ate only 400 calories of fat (44 grams), you'd still be eating 40% fat, considered a high-fat diet.

You could get your fat percentage down by eating more carbohydrate calories. Let's say you increased your total calories to 2000. Then you'd be on a "nice healthy 20% fat diet." But does that many any sense? Not to me.

Now let's look at what often happens when people go on a low carb diet. Let's say your previous lunch every day was a hamburger on a bun with french fries and a regular soda. I chose ground beef and a large hamburger bun (this had less fat than a fast-food burger, but I wanted to be able to compare it with a burger without the bun), a "serving" of Burger King fries, and a 32-ounce soda. Most people would probably also add catsup, or the fast-food burger would come with a sweet sauce, but I'm ignoring that. According to my nutritional program, without those extras, you have a meal with 1358 calories, 38 grams of fat, 10 grams of saturated fat, and 200 grams of carbohydrate.

Now let's say you go on a low-carb diet. You still eat at the hamburger place because all your friends do, but now you get a large burger without the bun. In place of the bun, you order a salad of mixed greens (I used 2 cups), with ranch dressing, and water or a sugarfree soda. This meal results in 324 calories, 15 grams of fat, 3 grams of saturated fat, and 4 grams of carbohydrate.

The first meal results in 25% fat (38 grams of fat x 9, divided by 1358), because of all the calories in the soda and the potato. The second meal results in 42% fat (15 x 9, divided by 324), because the total calories are so much lower that the fat makes up a larger proportion of the meal.

But would anyone claim that a meal that included french fries and a large soda (25% fat) was healthier than a meal that included salad greens and ranch dressing (42% fat)?

I don't think so.

This is why measuring a diet by the percentage of fat can be so misleading. This is why worrying about the fat content of low-carb diets can be so misleading.

Although my current low-carb diet includes about 60% fat, I don't think I'm eating any more fat than I used to eat. What I'm not eating is all the carbohydrate I used to put underneath that pat of butter or tablespoon of oil.

Biochemist Richard Feinman said all this in a more concise and more academic way when he wrote here, ". . . it is important to recognize that percentages are misleading. There are really three degrees of freedom in design or analysis of a weight loss experiment: two of the three macronutrients and the total calorie intake. It is unlikely that the percentage rather than the absolute amount of macronutrients is the controlling variable and at least three published studies show that carbohydrate reduction is not necessarily accompanied by replacement with either fat or protein but rather caloric reduction due to the carbohydrate removed."

In trying to unravel the very complex picture of the role of fat in human health, and for us its role in diabetes control, remember to scrutinize any research articles you read to see if the reseachers were measuring absolute amounts of nutrients or their percentages.

If only percentages, take the results with a grain of salt . . . or maybe a piece of cheese.

Also see how they determined the various nutritional intakes. Did they isoloate people in a ward and feed them carefully controlled meals? Or did they provide all the food on a take-out basis and trust that the participants weren't eating anything else? Or did they just provide "guidance" by a nutritionist who told them what kinds of foods they should be eating? Or did they just ask people to fill out food questionnaires after the fact?

You also need to see how the researchers defined their terms. They can differ a lot. For example, some people call a diet that has 45% carbohydrate (225 grams on a 2000-calorie diet) instead of 55 or 60% a low-carb diet and then claim that low-carb diets do this or that. Often this information cannot be found in the abstract of the article. You have to obtain and read the whole thing.

If you're reading an article about the effects of fat, you need to determine what other foods the subjects were eating. If you're on a low-carb diet, you'll burn a lot more fat than if you're eating both fats and carbs.

Unfortunately, the news media can't deal with these subtleties. They want interesting stories. And simplistic interpretations make for better stories. A story headlined "Whortleberries cure cancer" would get more readers than a story headlined "When fed a diet of 98% whortleberry, small percentage of highly inbred white mice see improvements in obscure cancer type that never affects humans ."

But you're smarter than the average reporter. So reader beware. Read and learn. But don't take any nutritional study as the last word. And especially, don't accept fuzzy fat words like "fat intake."