Atkins Vindicated? and Advice to Dieters
Obligatory Liability Line: I am not a doctor. This is not medical advice, it is a summary of what I know about diets and dieting. If you follow my advice and die, don't come crying to me.
Two articles in the NEJM (that's the New England Journal of Medicine) report that the Atkins diet does OK. All the Atkins fans are cheering. Do these articles vindicate Atkins? Probably not.
About the best that can be said for Atkins is that his diet seems to work, contrary to a strawman set of decade old expectations (no one today seriously thinks that the diet doesn't work, they doubt its wholesomeness and its long-term effect on health, which neither study investigated). Contrary to Atkins' teaching, the two articles in NEJM demonstrate that his reasoning for why his diet works is fundamentally flawed.
First, both studies showed what many (including yours truly) guessed early on: people actually eat fewer calories under Atkins than they do pre-diet, and in fact, fewer calories than you would on a conventional low-fat diet (people did not guess this second part, which is interesting). This is the only mechanism for weight loss. Ketosis is not necessary, nor is it helpful. The second study explicitly rules out ketosis as a factor in weight loss at 3, 6 and 12 months (no effect of being ketotic on weight lost), and in any case virtually none of the participants were ketotic at 6 months - probably showing that people's bodies accommodate to the lower carb levels. Also, critically, the reason for why people on a high-fat/high-protein/low-carb diet don't eat as many calories is unknown - an area for future research.
Also, in contradiction to the claim of Atkins and his fans, people on Atkins dropped out of the study (and the diet) with equal frequency to those on the conventional diet, belying the argument that Atkins is easier to adhere to. This suggests a different kind of advice for dieters - do whatever diet makes you happier and will allow you personally to cut calories in the long run.
Furthermore, at 12 months, the Atkins people were not statistically significantly different in weight loss than the conventional diet people (even when you only count those who didn't drop out of the study), largely b/c although both groups gained back weight from month 6 to month 12, the Atkins people gained it back faster than the non-Atkins people from month 6 to 12 (although directionally, the Atkins people were still better off, the trend suggests that 1.5 years out the two groups would have fully equalized).
As quoted in the discussion of the more relevant study ([1], see below this text for excerpt), it also turns out that evaluating the lipid profile of Atkins is hard because consuming all that fat tends to increase LDL, which is bad for your heart, but increase HDL and decrease Triglicerides, which is good for your heart. We don't know enough to know how to make that tradeoff (is raising HDL by 20% but also raising LDL by 7% and decreasing tri by 25% a good tradeoff? - sounds like it probably might be, but no guarantee).
Now, from a personal point of view the above is not too relevant, because the study was evaluating stand-alone Atkins, but not necessarily Atkins as many (such as people I know) follow. This is true because many people are both on Atkins as well as lipid-lowering drugs, and so decreasing triglycerides & increasing HDL may be a good strategy when LDL is being artificially held down by a pharmacologic agent (but if the drug also decreases HDL and LDL, then you'd have to do another study that included people on such meds to see if they get any benefit at all from Atkins). This is all good news, but by no means can one conclude that Atkins is "good" for your heart in an unqualified way.
Finally, the overall amounts of weight lost after both 6mos and 1 year were not impressive. Even if you take the maximum average weight loss (at 6 mos for both diets), people only lost, on average, 14 lbs on Atkins, 7 lbs on conventional. This is not a huge amount, considering that the average starting amount was 217 lbs. By the end of 1 year, average wt loss was about 2/3 of that, or 10 lbs in the optimistic case. This is not exactly an encouraging diet solution.
How to Diet
A reasonable summary of the state of our knowledge on dieting is the following:
One good way to eat fewer calories, as shown by these two studies, is to follow Atkins' advice. A reasonable summary of the evidence on Atkins is that Atkins is superior to other dieting methods if used as a short term weight loss regimen to show quick results and to drop the first bit of weight fast. If properly combined with enduring lifestyle changes, such as increased exercise and an overall sustained reduction in caloric intake (which may be facilitated by eating some fattier foods, which prove more filling), then this is probably the optimal strategy, because the results are so dramatic (about .7 lbs / week lost on Atkins on average for the first 6 months) and thus people can be encouraged to keep up their other lifestyle changes. Nonetheless, to the extent possible, the high-fat/low-carb diets should be rich in fish and other sources of non-saturated fat (such as nuts) rather than cream, butter, and beef. If you are diabetic, Atkins is the way to go.
Why do people on Atkins eat less? -- The mystery revealed
My explanation. I'll be vindicated soon.
The mechanism for why Atkins people consume fewer calories is due to a variety-seeking satiation trigger - animals satiate faster on homogeneous-flavors (such as many oriental dishes, risotto, etc) than on a variety of flavors. That's why we frequently can "save room" for dessert, even though the remainder of the main course is is unappetizing. Atkins drastically reduces the variety of foods available, and thus decreases appetite and caloric intake. For those geeks who want a citation, this was quoted in a recent article in a recent Journal of Consumer Research (Inman, 2001):
Research in the field of physiobehavior potentially explains why consumers might seek variety on some attributes more intensively than others. This literature has consistently reported that the pleasantness of a food just eaten drops significantly, and researchers have coined the term sensory-specific satiety to describe this phenomenon (e.g., Lyman 1989, Rolls et al. 1981b). Interestingly, the pleasantness of uneaten foods remains either unchanged or increases, and the decrease in preference for the eaten food appears to be enduring. Further, the underlying driver of the phenomenon seems to be more a function of the food's sensory characteristics than of its macronutrient composition.
[1] Excerpt from Article #2 in NEJM
The overall effect of the low-carbohydrate diet in comparison with a conventional diet on the risk of coronary heart disease in our subjects is uncertain. As compared with the conventional diet, the low-carbohydrate diet was associated with a greater improvement in some risk factors for coronary heart disease (serum triglycerides and serum HDL cholesterol), but not others (blood pressure, insulin sensitivity, and serum LDL cholesterol). Moreover, the clinical significance of the favorable changes in the HDL cholesterol triglyceride axis in the setting of a high fat intake is not clear.
[a long-winded and tortuous way to say that we don't know enough to say whether, on balance, the changes to peoples cholesterol levels were good or bad for them - EE]
