Abstract 2368: Baseline Dietary Glycemic Load as a Predictor of Weight Loss on the Atkins, Zone, Weight Watchers and Ornish Diets.
Background: Methods for matching patients with optimal diets are lacking. We previously reported that 4 popular diets (Atkins, Zone, Weight Watchers and Ornish), which range from low to high in dietary carbohydrate and glycemic load, produced similar 1-year weight loss in a randomized trial. We tested the hypothesis that dietary glycemic load at baseline (prior to dietary intervention) would affect the magnitude of weight loss from each diet.
Methods: We randomized 160 overweight and obese adults (mean BMI=35) to follow the Atkins, Zone, Weight Watchers, or Ornish dietary advice. Participants attempted maximum dietary adherence and attended 4 small-group classes during the first 2 months, then self-selected dietary adherence levels for the following 10 months. One-year weight loss for each diet was determined according to tertiles of baseline dietary glycemic load, assessed by 3-day food records, nutrition analysis software, and published glycemic index tables. Missing data were replaced with baseline values in an intent-to-treat analysis (n=40 per diet).
Results: The association between 1-year weight loss and dietary glycemic load at baseline differed significantly according to diet type (p<.01 for interaction). Participants in the highest tertile of dietary glycemic load at baseline had greater weight loss on lower carbohydrate diets (−4.6 kg, −4.6 kg, −2.1 kg, −1.5 kg for Atkins, Zone, Weight Watchers, and Ornish, respectively; p=0.05 for linear trend), and participants in the lowest tertile of dietary glycemic load at baseline had greater weight loss on higher carbohydrate diets (+0.3 kg, −1.0 kg, −3.2 kg, −5.3 kg for Atkins, Zone, Weight Watchers, and Ornish, respectively; p=0.01 for linear trend). For the middle tertile of baseline dietary glycemic load, weight loss did not differ between diets (−1.7 kg, −4.1 kg, −3.8 kg, −1.5 kg for Atkins, Zone, Weight Watchers, and Ornish, respectively, p=0.6 for linear trend, p=0.3 for two highest vs. two lowest).
Conclusions: Dietary glycemic load intake prior to dietary intervention may impact the effectiveness of popular diets for weight loss, and may be an important consideration in optimal diet selection for individual patients. These results should be verified in other study populations.