Three Diets Face Off

This study (the DIRECT study) has been getting a lot of attention. It appeared in the July 17 issue of the New England Journal of Medicine. It compared 3 diets over the course of 2 years: an Atkins-like diet, a Mediterranean (Med.) diet, and an American Heart Association (AHA) diet:

Weight Loss With A Low-Carbohydrate, Mediterranean, Or Low-Fat Diet (pdf), NEJM, 2008

In all groups, women were eating ~1500 calories/day, men ~1800 cal/day, or so they said. There was no difference among groups in amount of physical activity. The low-carb Atkins-like group had the highest drop-out rate, losing almost a quarter of participants over 2 years.

For weight loss, the Med. diet and the Atkins-like diet performed equally well -- when you looked at all participants.

Weight loss among all participants after 2 years: ¹

• AHA diet: 6 lbs
  
• Med. diet: 10 lbs
  
• Atkins-like diet: 10 lbs

But the Med. diet was the clear winner -- when you looked at just women.

Weight loss among women after 2 years:

• AHA diet: 0.2 lbs
  
• Med. diet: 14 lbs
  
• Atkins-like diet: 5 lbs

Here's something interesting, regarding diabetes ...
Go to the study here, flip to page 240 (12 of 13), and look at Graph D for Fasting Glucose.

The fasting glucose of people with diabetes eating the Med. diet decreased by an average of 32.8 mg/dl. That's astounding.

The fasting glucose of the people with diabetes eating the Atkins-like diet and the AHA diet didn't decrease at all. In fact, their blood sugar went up, even though they lost weight.

Look at Graphs E and F and you'll see something similar. People with diabetes eating a Med. diet had the greatest reduction in fasting insulin, and the greatest reduction in HOMA-IR (a measure of insulin resistance), meaning the cells of those eating the Med. diet had become more sensitive to insulin compared to the cells of those eating the other diets.

All this led the authors to conclude that, of the 3 diets tested, a Med. diet may be the most effective for glycemic control.

What's interesting about that is the Med. diet was high in carbohydrate.
________
What Was Unique About The Mediterranean Diet?

The study tells us:

• People eating it were consuming the highest amounts of dietary fiber.
  
• People eating it had the highest ratio of monounsaturated to saturated fat.

Other than that, there wasn't much difference, macronutriently-speaking, between what was called a low-fat diet (American Heart Association diet) and a Mediterranean diet. The percents of energy supplied by carbohydrate (50.7 vs. 50.2%), protein (19.0 vs. 18.8%), and fat (30.0 vs. 33.1%) were similar between those 2 groups. (The low-carb, Atkins-like diet was quite different, supplying 40.4% carbohydrate, 21.8% protein, and 39.1% fat.)

So, between the AHA and Med. diets, it comes down to foods. We know the Med. diet "was rich in vegetables and low in red meat, with poultry and fish replacing beef and lamb." There was a specific emphasis on olive oil (30 to 45 g/day) and nuts (5 to 7 nuts/day).

The AHA diet is described on their website. If it was anything like the AHA diet I reviewed here -- the one that allows frosted cupcakes, biscuits, French fries, hamburgers -- no wonder it lost out.
________¹ In addition to the weight losses listed here, which represent all 322 participants including dropouts, the authors also analyzed weight losses for only those who completed the study. In that analysis, the group with the highest drop-out rate lost the most weight. Those weight losses are being advanced erroneously. When you do not conduct an intention-to-treat analysis, that is, when you do not analyze all participants in all groups, whether they stuck with the study or not, you lose the effect of randomization. I think this is a good example of how different conclusions can be drawn from the same data.