Friday, May 29, 2009

The 1 Really Important Thing We Didn't Tell You About Losing Weight

The most interesting part of 10 Things You Need to Know about Losing Weight was the segment which focussed on an overweight actor. This lady revealed that she had always been large and that she felt that she had a 'slow metabolic rate'. She also stated that she had given up worrying up about her weight or trying to diet and just concentrated on eating a healthy diet and being active. Clearly, neither was helping her to lose weight.

This idea that you are overweight because you have a body that burns food off more slowly is, of course, a common belief, but it isn't the case. Indeed, as Gary Taubes explains in his book Good Calories, Bad Calories [Alfred A. Knopf, NY, 2007] in the chapter Paradoxes (p. 278):

The most obvious difficulty with the notion that a retarded metabolism ... is that it never had any evidence to support it. ... Magnus-Levy had reported that the metabolism of fat patients seemed to run as fast if not faster than anyone else's. .... The obese tend to expend more energy than lean people of comparable height, sex, and bone structure, which means their metabolism is typically burning off more calories rather than less. When people grow fat, their lean body mass also increases. They put on muscle and connective tissue and fat, and these will increase total metabolism...

For example, from Fitday, I, at 5'4" and 8 stone 3lbs, with a sedentary type of job require 2050 calories per day, whereas if I was 12 stone, I would need 2432 calories. As predicted, the metabolic rate for the subject of this segment came up perfectly normal.

Next came an investigation into how much she was eating. Looking at the lady in question, I guesstimated 3000 calories per day for her. She thought she was eating 1900 calories or less. As most people really have no idea about calories and portion sizes, I don't find this surprising and I don't think it's a deliberate or even an unwitting self-deception. It's just understandable, because humans didn't evolve doing complicated calorie counts before putting food into their mouths.

The test involved a 9-day diet record. Some of this was a video food diary plus a written food diary. The food that we saw looked perfectly reasonable: chicken and vegetables (though, my goodness - a whole head of broccoli?), (a very large) fruit salad, something dipped into a cup of coffee or tea. The guinea pig also drank doubly-labelled water so that the team could monitor her caloric intake. The result given to us was that she had underreported her food intake by 43% and that her actual intake was 3000 calories per day (score one me!)

Two things need commenting on here. Firstly, the underreporting. So what? It is fiendishly difficult to estimate portion sizes. I should know, I do it quite a lot to use Fitday. I try to be quite accurate, occasionally weighing or measuring to try and learn to 'eyeball' - say 30g of cheese or 1 oz of ham. On the other hand, there is a temptation to cheat. In my case, if I see the carbs going too high, there is a definite urge to downsize my estimates, even if this is silly because I'm only denying reality.

Secondly, what we weren't told. We weren't told what her energy expenditure was. We were told that the doubly-labelled water technique told the team that her caloric intake was 43% higher than her food diary records showed. The implied conclusion was that she was overeating - here we go, she thinks she's eating 2000 calories a day and she's actually eating 3000 calories a day and so she's fat. Whereas in fact that's probably not the case. She would probably be in energy balance - most people are, even fat people - they plateau at a certain weight, they don't all keep on getting fatter and fatter and fatter....

Now I thought this at the time, but I didn't know how right I was because then, to write this post I looked up the doubly-labelled water technique to see how it worked and found this :
the term doubly-labeled water test refers to a particular type of test of metabolic rate, in which average metabolic rate of an organism is measured over a period of time.
Energy expenditure measurements are easier to perform since the development and application of the doubly-labelled water technique.*
In other words, to work out that she was eating more calories than she said, they proved that she was using that many calories because the test itself measures energy expenditure not actual energy intake! But they didn't tell us that and they didn't draw the obvious conclusion that she was in energy balance! Certainly, if she ate less than 3000 calories a day she could draw on the stored fat and not starve. But, as she had pointed out herself at the start of the segment, she was not trying to diet and just trying to 'eat healthily'. Possibly to a dietician or nutritionist 'eating healthily' for an overweight person, by definition ought to mean dieting. However, the point is that just eating normally, she was not overeating, she was just eating as much as she needed to maintain her body - including all the extra fat - without discomfort or hunger.

Clearly, if she could only mobilise that fat and burn it up, she could eat less, so what is stopping that? The reason the fat mass isn't simply used as fuel as soon as we restrict calories is to do with the interplay of hormones in the body. In the simplest terms, if you have too much circulating insulin, then it promotes storage of fat in fat tissue and glucose burning in muscle. It does this because its job is to get excess glucose out of your blood because high blood sugar is damaging. Only if your insulin level is low, can your fat tissue release fat into the bloodstream and your muscles burn fatty acids.

The carbohydrate hypothesis of obesity basically says that carbohydrate intake promotes insulin release which promotes fat storage and the conversion of excess glucose (all starch breaks down into glucose) into fat for storage. So a person who eats a 'healthy diet' that would be 50-60% 'healthy' carbohydrates - 6-11 servings a day of cereals anyone? plus lots of fruit but is quite sensitive to this effect of insulin would convert all the excess to fat and store it. If they are unlucky their insulin stays relatively high preventing their body accessing this stored fat. Now they are 'growing' (but outwards) and as long as enough carbohydrates are consumed to keep insulin 'too high' for fat burning, their appetite tells them they need to eat more and so on it goes.

And that is why all nutritionists and dieticians and doctors should read Gary Taubes' book Good Calories, Bad Calories - published as The Diet Delusion in many countries.

*from Invited Commentary, Energy requirements assessed using the doubly-labelled water method, Klaas R. Westerterp, British Journal of Nutrition (1998), 80, 217–218. Available here.

Thursday, May 28, 2009

Review of 10 Things You Need to Know About Losing Weight

I managed to watch 10 Things You Need to Know About Losing Weight (BBC 1, 9pm) last night without throwing anything at the screen - click here for a link to the program (may be available in the UK only) or here for a related BBC Magazine article - but this may just have been because my hands were occupied taking notes so that I could blog about it.

Here are the 10 things:

  1. Not sure what they called this, seemed to be a variation on: Don't go shopping when you're hungry. The presenter had his brain MRI scanned, firstly on a full stomach, on a second occasion on an empty stomach, and while being shown pictures of food. His brain 'lit up' more in response to 'high calorie' food (chocolate eclairs) than 'low calorie' food (cucumber slices) when he was hungry; whereas when he wasn't hungry the response was the same.

  2. Use of an expensive machine to discover the bleeding obvious: when you are hungry, you are more interested in food and, more than that, more interested in food that your brain knows (from prior experience) is more likely to provide you with calories. Actually, I suppose this kind of research is necessary - we should investigate 'what everybody knows' - and try and disprove it - that is the scientific method. It also raises another interesting point - which wasn't mentioned in the program (although to be fair it isn't strictly relevant). Why doesn't the hungry brain respond to pictures of fruit and vegetables? Perhaps it's because humans didn't evolve chewing their way through bucketloads of plant matter for hours?

  3. Use plate size to trick the brain about portion size, i.e. use a smaller plate - less food will look like more.

  4. Personally, I think this works, I've used it with children the other way round - i.e. put the food on a larger plate and it looks like less so they eat it up. However, if you don't eat 'enough', you may end up hungry later on and snack!

  5. Count calories - save a few here and a few there e.g. have black coffee instead of cappuccino, and it will add up to fewer calories over the day/month/year, leading to you magically losing weight.

  6. Or maybe you'll just be hungrier? Has there ever been a controlled trial of this idea to see if it actually works?

  7. Fat people eat more than they think they do.

  8. Actually this was the most interesting point and worth a separate post of its own.

  9. Eat protein because it leads to greater satiety.

  10. This was demonstrated with a small-scale experiment during the programme. It has been shown by studies.

  11. Liquid food (as opposed to drinks with food) fill you up for longer. This was also demonstrated with an actual experiment.

  12. This point is different from the preceding one: the mechanism here is one of the physical constraints on digestion, whereas the fullness from protein comes from the release of a hormone (PPY) which interacts with other appetite regulating hormones (leptin, ghrelin). If you use soup to 'trick' your body into eating fewer calories than usual, you will likely eat more at the next meal, once the soup has been digested.

  13. Choice causes overeating.

  14. I thought the 'experiment' conducted to show this was quite poor. Two bowls of equal quantities of sweets were left out in an office canteen with a sign Free Sweets: one bowl was obviously smarties, the other bowl held purple smarties only. The smarties all disappeared, the purple ones didn't. A better comparison would have been smarties vs. chocolate buttons (they're all the same). The purple smarties looked vaguely medicinal - how do we know that didn't put people off?

  15. Calcium in dairy binds fat which you then excrete: over a month this can save you calories.

  16. In fact, it saved the guinea pig in this experiment slightly more than 5g of fat per day or about 160g per month. (Sounds like a lot? - I eat that much fat every day! I'm not going to quibble about the loss of one day's consumption per month.) Interesting - bad for dairy's image as a source of calcium - how much of the dairy calcium is lost this way? Also, the presenter felt constrained to recommend 'low-fat' dairy for this strategy. Noteworthy that this was the only vestige of 'low-fat' dogma in the programme. On the other hand, if it truly is 'low-fat' dairy then there isn't much point is there - where's it gonna find the fat to bind?

  17. Exercise - another interesting one. An experiment with the presenter on a treadmill showed that 90 minutes of fairly fast-paced walking (he was quite breathless after it) only burned about 19g of fat (171 calories).

  18. Given that after 90 minutes of fast-paced walking you've probably built up a bit of an appetite, it isn't going to do much for weight loss is it?
    However there is a punchline - an 'afterburn' effect where exercise boosts 'fat-burning' into the next day. So this could work - but there is still that question: why will the greater amount of calorie burning going on, not prompt your body to ask you for more food?
    Update: the effect of exercise to promote 'fat burning' is disputed by research - see article here.

  19. Small amounts of extra movement during the day e.g. take the stairs instead of the lift, will boost your calorie burning.

  20. Can't argue with this really, but the effect will be small (90 minutes on a treadmill = 171 calories) and once again it ignores the hormonal elephant in the room - which will be tackled in the next post about point 4.

Wednesday, May 6, 2009

Atherogenesis in Mice

Another day, another plug for the diet-heart hypothesis in BBC Health, even when the study being reported on Scientists pinpoint fats danger is really about molecular genetics(Thorp et al.,Reduced Apoptosis and Plaque Necrosis in Advanced Atherosclerotic Lesions ofApoe and Ldlr Mice Lacking CHOP, Cell Metabolism ,Volume 9, Issue 5, 474-481, 6 May 2009, subscription required). The study showed that mice lacking a gene (CHOP) which helps to trigger cell death (apoptosis) had a 35% smaller area of plaques and 35% less apoptosis and 50% less necrosis (dead tissue) in plaques. To quote the researhers directly (as reported by the BBC):

Lead researcher Dr Ira Tabas said that previous research had suggested that this mechanism might be involvedin plaque rupture, but the magnitude of the effect uncovered in the latest study was a surprise.
He said: "The fact that we were able to isolate one gene encoding one protein with such a profound effect on plaque necrosis (death) was a big surprise."
Dr Tabas said the finding raised hopes of new drugs which could act on the key gene, or the associated mechanism, to cut the risk of dangerous plaques.
"Just about everybody in our society has atherosclerosis (thickening of the arteries) by the time we reach 20," he said.
"So the wave of the future in treating atherosclerosis will be in preventing harmless lesions in young people from becoming dangerous ones, or soothing dangerous plaques so they don't rupture as we age."

Never mind what effect such a treatment might have on necessary cell death (e.g. to deal with emerging cancers) in other parts of the body.

Anyway, what does this have to do with diet and the heart? Well, again from the BBC article:
Scientists have identified a genetic mechanism which appears to determine which fatty deposits in the arteries have the potential to kill us. Most of these plaques pose no risk to health, but a minority burst, forming blood clots, which can cause heart attacks or strokes. .....
Fatty deposits begin to form in the arteries of most people in their teens, but the vast majority are harmless.

Here we see the perpetuation of the myth that fat just floats around in the bloodstream clogging up our arteries like it would a drainage pipe. Plaque formation is a much more complex process than that and its genesis is still not fully understood (see for example, extensive discussion here or here).

But, ah you say, just read on ...
The researchers bred mice prone to develop plaques, and fed them a high-fat diet for 10 weeks.

So what was this high-fat diet? It was the TD.88137 Western Diet (Teklad Lab Animal Diets, Harlan Laboratories, Madison, WI) which consists of:

Corn Starch150.0
Anhydrous Milkfat210.0
Mineral Mix, AIN-76 (170915)35.0
Calcium Carbonate4.0
Vitamin Mix, Teklad (40060)10.0
Ethoxyquin, antioxidant0.04

(Data from this pdf.)
This diet is 17.3% protein, 48.5% carbohydrate and 21.2% fat by weight, but 15.2% protein, 42.7% carbohydrate and 42.0% fat by energy, thus approximating a typical Western-style diet which is high in fat and simultaneously high in carbohydrate. Note that of the carbohydrates 70.4% by weight is sucrose! The mice are eating 30% of calories as sucrose. Now mice are not little people, but what does that kind of intake do to people?

How does this compare to a mouse's real diet?
From The Mouse in biomedical science (James G. Fox, Stephen W. Barthold, Muriel T. Davisson, Christian E. Newcomer, 2nd ed., Academic Press, 2007) p. 28 we learn that it is still debated whether mice are granivores, eating a wide range of cereals, oilseeds, and a variety of grass and plant seeds, or whether they live on a mix of plant and animal sources. However from the evidence presented in this book it appears that in many environments, mice eat small invertebrates for at least part of the year (i.e. when seeds are in short supply) or to supplement plant seed diets.

In short, the typical composition of a diet of invertebrates is high in fat and protein e.g. from p.41 in
Marsupial nutrition, (Ian D. Hume, Cambridge University Press, 1999) it can range from 20-60% fat and 10-75% protein (by weight of dry matter) for typical things that a mouse might eat (insects and insect larvae). Cereals are typically 68-79% carbohydrate, around 10-15% protein and 2-7% fat, legumes are as much as 25% protein, typically 50-60% carbohydrate and only 1-2% fat whereas nuts (e.g. hazelnut) and oilseeds (e.g. sunflower) are typically about 15-25% protein, 50-60% fat and 15-20% carbohydrate (from various tables in On Food and Cooking, H. McGee, 1st ed. Unwin Hyman, 1984).

From this we can conclude that a typical wild mouse would for part of the year eat a diet that was mainly protein and fat and for another part of the year eat a diet that was high in carbohydrate - at least if it ate cereals, but not so much if it ate other types of seeds - but low in fat. It would not however be eating a lot of sucrose. The carbohydrate in grains and seeds is starch which is a polymer of glucose and does not contain fructose. As further support of this analysis here Peter of Hyperlipid considered data on what wild-type mice eat when given free choice: about 12% protein, 6% carbohydrate and 82% fat (all as proportions of energy).

A final note about the mice. The mice used in the experiment were either apoe or ldlr mice. Apoe mice lack a particular lipoprotein (apolipoprotein E) which is important in both the HDL and vLDL cholesterol transporters, in particular:
ApoE mediates high affinity binding of chylomicrons and vLDL particles to the LDL receptor, allowing for specific uptake of these particles by the liver, preventing the accumulation of cholesterol rich particles in the plasma
Mice develop normally, but exhibit five times normal serum plasma cholesterol and spontaneous atherosclerotic lesions
Ldlr mice lack a proper LDL receptor and essentially mimic (familial) hypercholesterolaemia with a very high circulating LDL level and an increased propensity to develop atherosclerotic lesions amongst other things.

Does this not indicate that fat is the root cause? Well not necessarily.

vLDL is made in the liver to transport triglycerides (made from excess carbohydrates intake) to the tissues for use and storage. At this stage, it does not contain apoE: it has to pick that up from HDL on the way. ApoE contributes to its recognition and re-uptake by the liver after it has performed its delivery task or it loses its apoE and becomes an LDL particle and is taken up by body cells with an LDL receptor. So, this process will become disrupted in an apoe mouse which does not have a proper apoE protein. No wonder it ends up with excess blood cholesterol (which really means excess circulating lipoproteins). Similarly as ldlr mice lack the LDL receptor, they cannot remove the LDLs left at the end of the described process. On the other hand, after digestion, fat is absorbed either directly into the bloodstream - if the molecule is small - which gets it to the liver (where it may contribute to triglyceride production) or, for larger molecules, as chylomicrons which go via the lymphatic system into the bloodstream and from there directly to fat tissue for storage or to the liver to be used to provide fuel.
So which is more likely to contribute to the problem - the 21.2% of food (by weight) that comes as fat (most of which doesn't go straight to the liver anyway) or the 48.5% of food (by weight) that comes as carbohydrate - three-quarters of which is sucrose and half of that is fructose (i.e. 17% by weight of the total food intake) which goes straight to the liver and comes out as triglycerides.