Letters Section Editor: Jody W. Zylke, MD, Senior Editor.
Author Affiliations: New Balance Foundation Obesity Prevention Center, Boston, Massachusetts (Drs Ludwig and Ebbeling) (firstname.lastname@example.org); and Clinical Research Center, Boston Children's Hospital, Boston, Massachusetts (Dr Feldman).
In Reply: Dr Weaver correctly observes that the trend test with equal spacing among diets from lowest to highest glycemic load (low-fat, low–glycemic index, and very low-carbohydrate diet) amounts to a pairwise comparison of the low-fat and very low-carbohydrate diets. We obtained an identical result, which was not reported, when the diets were spaced according to calculated glycemic load (185.1, 51.1, and 3.9 g/2000 kcal/d, respectively; P = .05). In any event, we subjected all outcomes to a 3-group analysis of variance and to a test for linear trend. We agree that the former is more stringent than the latter, and that interpretation of the CRP result requires caution. Some1,2 but not all3 studies of this topic have similarly reported higher CRP levels for a low-carbohydrate vs low-fat diet, and this issue warrants additional study.
Ludwig DS, Ebbeling CB, Feldman HA. Dietary Composition During Weight-Loss Maintenance—Reply. JAMA. 2012;308(11):1087. doi:10.1001/2012.jama.11620