Title and Abstract
Gardner CD et. al. Comparison of the Atkins, Zone, Ornish, and LEARN diets for change in weight and related risk factors among overweight premenopausal women: the A TO Z Weight Loss Study: a randomized trial. JAMA. (2007) 297(9):969-77.
CONTEXT: Popular diets, particularly those low in carbohydrates, have challenged current recommendations advising a low-fat, high-carbohydrate diet for weight loss. Potential benefits and risks have not been tested adequately. OBJECTIVE: To compare 4 weight-loss diets representing a spectrum of low to high carbohydrate intake for effects on weight loss and related metabolic variables. DESIGN, SETTING, AND PARTICIPANTS: Twelve-month randomized trial conducted in the United States from February 2003 to October 2005 among 311 free-living, overweight/obese (body mass index, 27-40) nondiabetic, premenopausal women. INTERVENTION: Participants were randomly assigned to follow the Atkins (n = 77), Zone (n = 79), LEARN (n = 79), or Ornish (n = 76) diets and received weekly instruction for 2 months, then an additional 10-month follow-up. MAIN OUTCOME MEASURES: Weight loss at 12 months was the primary outcome. Secondary outcomes included lipid profile (low-density lipoprotein, high-density lipoprotein, and non-high-density lipoprotein cholesterol, and triglyceride levels), percentage of body fat, waist-hip ratio, fasting insulin and glucose levels, and blood pressure. Outcomes were assessed at months 0, 2, 6, and 12. The Tukey studentized range test was used to adjust for multiple testing. RESULTS: Weight loss was greater for women in the Atkins diet group compared with the other diet groups at 12 months, and mean 12-month weight loss was significantly different between the Atkins and Zone diets (P<.05).
Mean 12-month weight loss was as follows: Atkins, -4.7 kg (95% confidence interval [CI], -6.3 to -3.1 kg), Zone, -1.6 kg (95% CI, -2.8 to -0.4 kg), LEARN, -2.6 kg (-3.8 to -1.3 kg), and Ornish, -2.2 kg (-3.6 to -0.8 kg). Weight loss was not statistically different among the Zone, LEARN, and Ornish groups. At 12 months, secondary outcomes for the Atkins group were comparable with or more favorable than the other diet groups. CONCLUSIONS: In this study, premenopausal overweight and obese women assigned to follow the Atkins diet, which had the lowest carbohydrate intake, lost more weight and experienced more favorable overall metabolic effects at 12 months than women assigned to follow the Zone, Ornish, or LEARN diets. While questions remain about long-term effects and mechanisms, a low-carbohydrate, high-protein, high-fat diet may be considered a feasible alternative recommendation for weight loss.
When this paper came out a couple of years ago, there was a ton of press about this study with further claims of low-carb metabolic advantages (several fitness related blogs have already stated that the Atkins diet generated significantly greater weight losses) based on it; of course as you’ll see the claims that were made based on the results aren’t quite as astounding as made them out to be.
I want to point out up front that I am hardly against low-carb diets even though my comments about them often leads people to think that especially when people start talking about ‘metabolic advantages’ of such diets. For goodness sake, my first book The Ketogenic Diet was dedicated to nothing but low-carbohydrate diets and many of my diet books incorporate some type of carbohydrate restriction.
However, while I like low-carb diets and think that they are appropriate under some (but assuredly not all) situations, I don’t believe that the research supports much of a metabolic advantage (in terms of being able to lose more weight/fat at the same or higher caloric intakes). If such diets have an advantage in terms of dieting, it tends to have more to do with adherence and food intake due to appetite suppression. Which is still a benefit, mind you. But is different than what is often being claimed.
I would entreat people to reread the above paragraph two or three more times before they start entering comments about how I’m anti-lowcarb diets.
Which brings us to this week’s study. The researchers set out to compare 4 diets of drastically different carbohydrate intake. The first was Atkins which is a very low-carbohydrate diet. The second was the Zone which is a moderate carb diet (40% carbs, 30% protein, 30% fat). The third is something called the LEARN diet which is your basic food Pyramidy type of diet with 55-60% carbs and saturated fat below 10% of total calories. Finally was Dean Ornish’s extremely high-carbohydrate, very low-fat (10% or less) diet. These diets were chosen to represent the spectrum of diets from very low carbohydrate to very high carbohydrate.
The subjects were premenopausal women between the ages of 25-50 with a body mass index between 27 and 40 who had been weight stable for the previous 2 months. Folks were excluded for various reasons. 311 subjects entered the study and were randomly assigned to one of the four groups with about 70 subjects per group.
All subjects were given a copy of the respective book and a dietitian explained the details of the diet to each. For the first 8 weeks of the study, all subjects attended a 1 hour class per week. For the remaining 10 months of the study, they were on their own.
Subjects did receive email and phone contact from the staff between the 2 and 6 month mark and the 6 and 12 month mark and small financial incentives were given for completing the data collection at the 2, 6 and 12 month time point.
Diet was assessed by a 3 day food recall (I’ll come back to this below) and energy expenditure was estimated by a 7 day activity recall. Subjects were measured for height, weight, body fat was done by DEXA. A number of blood measures including total cholesterol and blood triglycerides were measured. So was fasting insulin and blood glucose along with blood pressure.
Ok, before getting into the details, I want to look at the overall results since that’s most of what people focused on. After 12 months on the diet, the respective weight losses were
- Atkins: 4.7 kg (10.3 lbs)
- Zone: 1.6 kg (3.5 lbs)
- LEARN: 2.6 kg (5.7 lbs)
- Ornish: 2.2 kg (4.8 lbs)
So yes, the Atkins group did get better results, 2.5 kg or more weight loss than the other diets over the span of a year. And, according to the self-reported food intakes (an issue that I’ll discuss momentarily), they did it eating the same number of calories as the other groups with both groups reporting a reduction in food intake over the length of the study. AHA, more weight loss on the same calories, there’s your metabolic advantage….
Ok, first and foremost, let’s be realistic: regardless of the fact that Atkins got double the weight loss, those results suck. Ten pounds weight loss in one year amounts to a 3/4 lb weight loss per month in the Atkins group and half or less than that in the other groups. By contrast, low-calorie diets that are highly controlled can generate a 7kg/15 lb weight loss over 4 weeks. My own Rapid Fat Loss Handbook can do that in 2 weeks in some people.
Yes, fine, the study points out that even small weight losses can improve health but what dieter would be happy with that? Not many. More like 10 lbs weight loss in 2 months. I didn’t pick that value out of a hat.
The paper shows changes in body weight at each of the 2, 6 and 12 month time spans. And Table 3 in the paper, which I show below shows how each diet affected weight. It tells the entire tale so far as I’m concerned.
There are a few key observations to make from this. The first is this: at the 2 month mark, the Atkins group was already 4kg down in body weight while the other groups had lost about 2.5 kg or so. Recall from above that the total difference in weight loss between Atkins and the other groups was only about 2kg. So most of the difference between the diets occurred in the first 2 months.
Now, it’s well-established that ketogenic diets can cause significant water loss in the first couple of weeks, water loss can range from 1-15 lbs over that time frame. I’m a little guy but I can drop 7 lbs in 3 days of carbohydrate restriction or about 2.5 kg. So the 2 kg ‘advantage’ of the Atkins diet not only could be due to water loss but in all likelihood is due to water loss. Other studies, lasting from 4 days to 2 weeks show the same 2kg difference in weight loss. All of which occurs early on and likely represents water drops due to carbohydrate restriction.
Another interesting point is that over the next 10 months, the Zone, LEARN and Ornish group didn’t lose an additional pound and even showed a slight trend towards regain. Read that and let it sink in for a few minutes. Over a 12 month diet, after a small weight loss in the first 2 months, there was no additional weight loss for the next 10 months.
Rather, the entirety of their weight loss occurred during the first 8 weeks when they went to weekly meetings and they didn’t lose an ounce for the remaining 10 months. In contrast, the Atkins group had lost about 2 kg more at the 6 month mark and regained over a kilogram at the 12 month mark, leading to the final results reported above.
As mentioned, the groups all self-reported eating roughly the same number of calories and reducing their caloric intake over the length of the study. But let’s think about that rationally for a second: are we to honestly believe that the three groups which didn’t lose an additional pound over 10 months truly ate less over that time period? Was every single person in this study one of the metabolic miracles that exist in droves on the Internet, that can eat less let magically maintain weight?
Or is it more likely, as with tons of studies done previously, that they were under-reporting their food intake and actually eating more (possibly quite a bit more as studies of carb based diets show a systematic under-reporting of anywhere from 30-50%) than they thought or said they were?
Now, call me a cynic but I think you know where my opinion on this lays. As I discuss in the research review Ketogenic Low-Carbohydrate Diets have no Metabolic Advantage over Nonketogenic Low-Carbohydrate Diets, when calories are strictly controlled (and protein intake is identical), there is simply no metabolic advantage or greater fat loss to be had. Another study, which I will eventually review for the site was unable to find any measurable difference in metabolic rate for ketogenic vs. carb-based diets. As well, caloric misreporting on carb-based diets is known to be prevalent and the only logical answer to the claims of this study (e.g. the Atkins dieters lost more weight despite ‘eating the same amount’) is that the self-reported food intakes are invalid (as they usually are).
So even was slightly greater for the Atkins diet, it’s not because of any inherent metabolic advantage. It’s because, under uncontrolled conditions, people on ketogenic diets typically eat less. In this study, they only ate a little bit less because they only lost about 2 lbs over the 10 months of the uncontrolled study period.
It’s worth noting that the researchers point out that the weight loss trajectories, meaning the trends over time, indicate that the Atkins dieters were regaining weight. So over a longer time period for the study, the differences in weight loss would have been even smaller between the Atkins group and the others.
I should mention that based on the self-reported food intakes, a lot of the criticism of this study has to do with overall compliance to the actual diets. For example, Ornish has complained that the dieters were eating 30% fat when he prescribes only 10% fat; hence it wasn’t his diet. And the Atkins dieters were eating almost 35% carbs at the end of the study. So it wasn’t really an Atkins diet.
There is some truth to this and does raise some questions about the inherent validity of the study. However, it also raises the point that, in free living subjects, people usually suck at adhering or properly following diets. So even with those flaws, this study probably represents how people actually diet in the real world.
I should mention that some of the health measures (blood lipids, etc) did show a slight advantage to one or the other diets but the differences were small and hardly significant. You could generate more changes with just about any reasonable diet that actually too a decent amount of weight/fat off of someone than with the pitiful results this study found.
First, what I’m not saying. As usual, folks will find a way to read this as “Lyle dislikes low-carbohydrate diets” which is incorrect. Like all dietary approaches they have their pros and cons and are appropriate in some conditions and not in others.
For many people, they make controlling calories easier, for people with insulin resistance they often improve health parameters to a greater degree than carb-based diets. Those are advantages to be sure but they aren’t the ones that most are fixated on (e.g. the idea that you’ll somehow lose tons of weight and fat while ignoring caloric intake).
Clearly, despite some of the current claims, simply reducing carbs doesn’t magically ‘cure’ obesity if calories don’t come down. And studies like this demonstrate that. Even if the Atkins diet was slightly superior to the other diets, the simply fact is that the overall weight loss was minor in all groups (this is a common finding among many studies where caloric restriction isn’t put into place).
Ten pounds total weight loss (or even true fat loss) in a year of dieting is crap results, plain and simple; whether or not it was slightly better than the other approahces wouldn’t seem to be that massively significant. Especially when you consider that 8 of those 10 pounds occurred in the first 2 months, meaning that there was only 2 lbs more net weight loss over the next ten months of the study (the equivalent of 0.2 lbs weight loss per month).
- Ketogenic Low-Carbohydrate Diets have no Metabolic Advantage over Nonketogenic Low-Carbohydrate Diets – Research Review
- Ketogenic Diets: High-fat or High-Protein – Q&A
- Reduced Fat and Reduced Carbohydrate Diets and Fat Loss – Research Review
- All Diets Work: Qualification
- Size of Deficit and Muscle Catabolism – Q&A