Clinical Question: Is a low-carbohydrate diet more effective than a low-fat diet at promoting weight loss?

Setting: Outpatient (any)

Study Design: Randomized controlled trial (nonblinded)

Synopsis: Low-carbohydrate (e.g., Atkins) diets are very popular, but are they really effective? This is one of two studies published in the same issue that attempt to answer this question but raise many new questions in the process. Researchers at a Veterans Affairs medical center identified a group of very obese outpatients (mostly men) with a body mass index (BMI) of at least 35; 58 percent of the patients were black, and 40 percent had diabetes. The mean BMI was 41.

Patients were randomly assigned (allocation concealed) to a low-fat diet consistent with the recommendations of the National Heart, Lung, and Blood Institute or a low-carbohydrate diet that recommended consuming less than 30 g of carbohydrates per day and choosing fruits and vegetables with a high ratio of fiber to carbohydrates. Dietary instructions were given through two-hour training sessions held weekly for four weeks, followed by one-hour sessions held monthly for another five months. The primary outcome was weight loss at six months, but many participants were lost along the way—only 53 percent were evaluable at six months. This is an important limitation of the study.

The last weight was carried forward as a proxy for the six-month weight in those lost to follow-up, which is appropriate. Patients on the low-carbohydrate diet lost more weight (5.7 kg [12 lb, 9 oz] versus 1.8 kg [3 lb, 15 oz] for patients on the low-fat diet; P = .002). A substantial weight loss (at least 10 percent of body weight) occurred in 14 percent of low-carbohydrate dieters compared with 3 percent of the low-fat dieters. White patients lost more weight than black patients (13 kg versus 5 kg; P = .009).

Interestingly, a greater decrease in triglyceride levels and a trend toward a greater decrease in glycohemoglobin levels were seen in subjects on the low-carbohydrate diet, but not in those on the low-fat diet. There was no change in the levels of total cholesterol, high-density lipoprotein cholesterol, or low-density lipoprotein cholesterol in subjects on either diet. One patient in the low-carbohydrate group died from complications of hyperosmolar coma, but the authors do not believe that the diet was responsible.

A second study found largely similar results at six months (Foster GD, et al. A randomized trial of a low-carbohydrate diet for obesity. N Engl J Med May 22, 2003;348:2082–90). However, this smaller trial, which also had a large dropout rate, found that the differences between groups became smaller and often nonsignificant after one year.

Bottom Line: This small study with a high dropout rate found that a low-carbohydrate diet can help patients lose more weight in the short term than a conventional low-fat diet. Although a low-carbohydrate diet does not cause adverse changes in lipid levels, it may not help patients live longer, better lives, which is the goal that should concern us most. Also of concern is the diminishing benefit after 12 months. (Level of Evidence: 2b–)

Used with permission from Ebell M. Low-carb (Atkins) diets help you lose weight. Retrieved July 24, 2003, from: www.InfoPOEMs.com.