Maybe Atkins Was Right

Abstract & Commentary

By Barbara A. Phillips, MD, MSPH

Professor of Medicine, University of Kentucky; Director, Sleep Disorders Center, Samaritan Hospital, Lexington

Dr. Phillips serves on the speakers bureau for PotomaCME.

Synopsis: Compared with a low-fat diet, low-glycemic index and very low-carbohydrate diets resulted in higher resting and total energy expenditure (thus more calories burned) during the weight loss maintenance phase after a low-calorie diet.

Source: Ebbeling CB, et al. Effects of dietary composition on energy expenditure during weight loss maintenance. JAMA 2012;307:2627-2634.

THIS STUDY OF 21 YOUNG ADULTS INCLUDED RUN-IN AND TEST PHASES. THE run-in phase included collection of baseline data, calorie restriction to achieve a 12.5% decrease in body weight, and determination of energy (calorie) requirements necessary to maintain weight at the reduced level. During the test phase, the investigators used a three-way crossover design to evaluate test diets (low-fat, low-glycemic index, and very low-carbohydrate) in random order; thus, every subject experienced each of the three diets.

The run-in (weight loss) diet was consistent with the Acceptable Macronutrient Distribution Range specified by the Institute of Medicine.1 The low-fat diet was designed to reflect current conventional recommendations to reduce dietary fat, emphasize whole grain products, and include a variety of vegetables and fruits.2 The low-glycemic index diet replaced some grain products and starchy vegetables with sources of healthful fat and low-glycemic index vegetables, legumes, and fruits. The low-fat and low-glycemic index diets had similar protein and fiber contents. The very low-carbohydrate diet was modeled on the Atkins diet and had severe restriction of carbohydrates. Participants took supplemental fiber with the very low-carbohydrate diet and also took multivitamins and mineral supplements throughout.

Measurements included resting energy expenditure (REE), total energy expenditue (TEE) as well as leptin, thyroid stimulating hormone, triiodothyronine, free urinary cortisol, insulin sensitivity, high-density lipoprotein (HDL) cholesterol, total cholesterol, triglycerides, plasminogen activator inhibitor 1 activity, C-reactive protein (CRP), blood pressure, participant ratings of hunger and well-being, and physical activity.

Although the investigators enrolled 32 participants, only 21 finished the study. This final cohort included 13 men who had a mean body mass index (BMI) of 34.4 kg/m2 and a mean age of 30 years. During the run-in (diet) phase, participants lost a mean of 31.5 pounds (14.3 kg), corresponding to 13.6% of baseline body weight. This phase included 12 weeks of weight loss and 4 weeks of weight stabilization. Percentage body fat decreased from a mean of 33.6% at baseline to 29.1% after weight loss. Mean calorie intake during the test diet phase was 2626 kcal/d.

The participants then underwent 12 weeks of weight maintenance, spending 4 weeks on each of the diets (low-fat, low-carbohydrate, low-glycemic index). As predicted, calories burned were reduced overall for all dietary conditions, a usual response to weight loss.3 But calories burned during weight-loss maintenance differed significantly among the three diets. The decrease in REE from pre-weight-loss levels was greatest for the low-fat diet (mean, -205 kcal/d), intermediate with the low-glycemic index diet (-166 kcal/d), and least for the very low-carbohydrate diet (-138 kcal/d; overall P = 0.03). In other words, people who ate the low carbohydrate/high fat (Atkins-like) diet burned more calories during the weight-maintenance phase than those in the other two groups. There was a similar pattern in TEE.

Serum leptin was highest with the low-fat diet, intermediate with the low-glycemic index diet, and lowest with the very low-carbohydrate diet. Cortisol excretion was highest for the very low-carbohydrate diet. Indexes of peripheral (P = 0.02) and hepatic (P = 0.03) insulin sensitivity were lowest with the low-fat diet. Measures of lipids were most favorable with the very low-carbohydrate diet and least favorable with the low-fat diet. However, CRP tended to be higher with the very low-carbohydrate diet. Blood pressure did not differ among the three diets. Ratings of subjective hunger and well-being were not different between diets. And actual weight change did not differ among the three diets.


Anyone can lose 20 pounds. The trick is to keep it off. This study helps us understand why “low-fat” diets might not be the best recommendation for individuals who are struggling to maintain weight after weight loss. In this rigorous trial, a small number of people who had already lost significant weight burned calories at a higher rate when on a low-carbohydrate (high-fat, Atkins-like) diet than on low-fat or low-glycemic diets. They also had greater improvement of most measures of the metabolic syndrome. Further, they experienced changes in leptin that are associated with improved success at weight loss.4 Opposite changes were observed with the low-fat diet, with the low-glycemic load diet being somewhere in between. The implication is that low-fat diets, which result in greater reduction in calories burned, may increase the likelihood of regaining weight lost in those who are dieting. In their commentary, the authors note that, “The results of our study challenge the notion that a calorie is a calorie from a metabolic perspective.”

This article received a lot of attention in the lay press, as Americans — two-thirds of whom are overweight or obese5 — continually search for secrets to weight loss. But there are several cautions and caveats here. First, there were no differences in actual weight-loss maintenance in the three groups in this short-term (4 weeks on each diet) study. Second, there were two negative outcomes observed during the low-carbohydrate diet: increased urinary cortisol (a hormonal measure of stress) and a tendency for increased CRP. These two factors are associated with increased cardiovascular risk. Because of this, the authors noted, “These findings suggest that a strategy to reduce glycemic load rather than dietary fat may be advantageous for weight-loss maintenance and cardiovascular disease prevention.”

It is worth noting that although the investigators enrolled 32 participants, only 21 finished the study, despite the benefits of a well-controlled weight loss program and a financial incentive ($500). Dieting is hard! In fact, only one in six overweight or obese adults report ever having maintained weight loss of at least 10% for a year.5

So what does this mean for our patients? At the very least, these results challenge the conventional dogma that low-fat diets are the only way to go. As with many other things in medicine, a “one size fits all” approach is unlikely to be effective for everyone. Ongoing work6 indicates that tailored treatment approaches might improve our rather pathetic attempts to address the chronic illness known as obesity. Patients might need to be encouraged to try different things until they find something that works.


1. Institute of Medicine. Dietary Reference Intakes for Energy, Carbohydrate, Fiber, Fat, Fatty Acids, Cholesterol, Protein, and Amino Acids. Washington, DC: National Academies Press; 2002.

2. Klein S, et al. Am J Clin Nutr 2004;80:257-263.

3. Sumithran P, et al. N Engl J Med 2011;365:1597-1604.

4. Crujeiras AB, et al. J Clin Endocrinol Metab 2010;95: 5037-5044.

5. Flegal KM, et al. JAMA 2010;303:235-241.

6. Ogden LG, et al. Obesity (Silver Spring); 2012. doi: 10.1038/oby.2012.79. [Epub ahead of print.]