Thursday, March 15, 2007

Diets Are Hard

Dr. Christopher Gardner is one busy researcher.

Remember my post last week about the garlic study? The one that received lots of media attention for its failure to show garlic useful in reducing cholesterol? They used a prepared garlic mash that had been previously frozen and thawed. Maybe fresh would have been different, but the study would have been difficult to conduct and control without some sort of monitored dosing.

Well, The lead author on the garlic study was Dr. Christopher D. Gardner of Stanford University. And the lead author on the study below, published a week later, this time in JAMA (March 6th issue) is Dr. Christopher D. Gardner of Stanford University:

Comparison of the Atkins, Zone, Ornish, and LEARN Diets for Change in Weight and Related Risk Factors Among Overweight Premenopausal Women

So here we have it. A straight-up comparison of 4 of the most popular diets, representing a spectrum of carbohydrate intake.

The study lasted for a year and was conducted as sanely and scientifically as any good study would be - using hundreds of women, hours of classroom instruction, books, handouts, incentive emails and phone calls, regular blood collections, regular dietary intakes ... incentive payments! Participants had an average education level beyond 4 years of college. Apart from locking these women up and rationing their food, this protocol, using this group of well-educated, motivated, financially- and socially-supported women, should have spelled dietary success.

Here are these women's weight losses:

The average weight loss across diets was 6 lbs, in one year. That's 1/2 pound per month.1 The women selected for this study were required to be overweight or obese at baseline. So, it's not 6 lbs from an already slender frame.

Note that after 6 months, all groups started to regain weight lost. It would be interesting to see what their weights were a year after the study, after all those incentives ended.

The Winner?

It looks like Atkins edged out the other 3 diets when it came to weight loss. Atkins' women lost an average of 10 lbs in 12 months, compared to a combined average of 5 lbs for the other 3 groups. I would call that only winning by a nose, especially after a year of dieting. But a nose is a nose. And had women in the Atkins group actually been eating the Atkins diet, that nose might have been a full head length.

What Were Participants Actually Eating?

I created these charts from data in the study, a chart for each macronutrient: carbohydrate, protein, and fat. Diets often differ by macronutrient content. Gardner, et al., selected these diets based on their range of carbohydrates, from Atkins (low) to Ornish (high).2

Before you run out to purchase Dr. Atkins New Diet Revolution and start restricting your carbohydrates to less than 10% of calories, take a look at what these women were actually eating at 12 months (even with all that diet instruction and prodding) compared to what they were instructed to eat (their goal) in each diet category. It may be, as Gardner concluded, that for those who "initiate a low-carbohydrate diet ... weight loss is likely to be at least as large for any dietary pattern."

Two comments ... Atkins is not so much a high-protein diet as it is a high-fat diet. With all that fat, participants still lost weight and their lipids were better (more about that in a follow-up post.)

Also, notice how all the diets' carbohydrate, protein, and fat contents converged or sought a middle ground over time. If the diet's prescription was too extreme (e.g. very low fat: Ornish, or very low carb: Atkins), participants soon adjusted that to a middle ground. It's as if we naturally crave a certain proportion of carbohydrate, protein, and fat. Or at least that's what's easy, available, and tastes good.

One last thing, regardless of macronutrient content, all of the women had reduced their calorie intake by about 300 calories/day (at 12 months).

...Continued in "Fear Not The Fat".

1 offers one critique of why the women in this study didn't lose more weight:
In Year-Long Study, Atkins Edges Out Competition
In a nutshell ... people who are insulin resistant (prediabetic) do better on a low carbohydrate diet, while people who are insulin sensitive do better on a low-fat diet. The study's random assignments resulted in mismatches.

However, this is exactly what randomization is designed to overcome, and does if a study has a large enough population. You can always cherry-pick the most sensitive individuals to receive an intervention, and show the intervention good. But how does the intervention fare in the real, non-cherry-picked world? The answer to that question lies in this study's findings.

2 The LEARN diet (Lifestyle, Exercise, Attitudes, Relationships, and Nutrition) is based on the USDA's Food Guide Pyramid.

Cartoon via

No comments: