More Controversy!

Here's Blakes analysis of the LA Times article I put in my last post. How are we to know what to think?:


Hey Tess... here's one of the articles alluded to in the LA times column. It's pretty interesting to look. You're probably not interested in these details, though, and if not I'm sorry, but that's where the devil is. It shows how much each study has to be stared at to make sense, and how hard it can be to make sense of unrelated studies.

I guess if nothing else, as a friend, this shows you what a drag it is to be me.

Don't bother to read on if you don't want to see my analysis...

First, the study includes a total of 40 people. That is a very small number.

Second, it's important to segregate what they say about weight loss, fats circulating in the blood, and actual adipose tissue.

The weight before was about 210 pounds (plus/minus 20 pounds) for both grups. The high fat diet end weight was 190 (plus/minus 20 pounds) while the carb diet end weight was 195 (plus/minus 20 pounds). So, really, this study isn't making a claim about weight; the errors make the differences between and before/after indistinguishable. A better number to report would probably have been the % weight loss. This might wash out the big errors (comparing someone that was 170 with someone 210 won't reveal how each individual fared). Of course, this may have looked even worse than body mass.

The body fat was actually less in the high carb group.... 36.8% vs 38.2%... though the size of their error bars (8%) actually dwarfs any difference. The carb group had more non-fat body mass, which makes them slightly heavier overall.

Where they do see something striking is in concentrations of blood lipids. A decrease in triacylglycerols of 50% (as he said in the article). The carb group also saw a decrease of 20% (not mentioned in the article), which is also really striking. The error bars in both these numbers is 25%. Caloric reduction in general is a good thing for diabetes risk.

The other thing that's striking is the shift in the insulin response. This gets to that homeostasis thing I was mentioning before... the reduced carbohydrate load allows the body to shift to a lower insulin response: there's less sugar to import, so we require a smaller insulin response. On the other hand, the difference is response is really not too surprising: the body responds to the stimulus... less carbs, less insulin required to move fewer carbs into the cell. I bet if they looked after week1 of the study, they'd have seen the same thing.

To be fair, I think the metrics of diabetes risk is what they're emphasizing (circulating lipids). So, you can still be overweight (if I lose 20 pounds, I'll still be overweight), but reduce your risk for diabetes. Which is more important? I don't know. Being overweight but diabetes free is still probably linked to self-esteem and depression issues.

So, I guess the bottom line is that it's easy to confuse weight loss and health benefits. Also, such small studies are very hard to interpret because the errors are so large. However, it is nice to know that, even if you're too fat (at 210 myself, I can say that's too fat), you can still avoid risk for diabetes!

In the end, for me, the article is about how 210 pound people go on 2000 calorie diets for 12 weeks. I wish I could do that with a house full of pumpkin pies, candied yams, mashed potatoes, fudge and rolls.

Anyway, sorry! But thanks for posting the link, it was useful and fun for me!

bg