The review by Denke1 paints an optimistic picture of the efficacy of cholesterol-lowering diets in free-living subjects. We have also reviewed the literature in this area and believe that her summary of the evidence may overestimate the effectiveness of dietary intervention. A defined strategy for the identification of data in review articles is important to minimize citation bias.2 It is not clear how articles were selected for inclusion in Denke's review, and we found that there are relevant studies that were not reviewed, including some in which the effects of dietary advice on cholesterol were less impressive than those cited. For example, in a randomized trial that included 437 patients in Oxfordshire, England, Baron and colleagues3 reported that advice to lower total fat in the diet led to a reduction in cholesterol of less than 1% at 1 year.
In addition, the precision of individual estimates was