Low rates of dietary adherence and modest reductions in body weight and cardiac risk factors were found among participants in all four diets. Mean weight loss at one year was 2.1 kg for Atkins (21 [53 percent] of 40 participants completed; P=0.007), 3.0 kg for Weight Watchers (26 [65 percent] of 40 completed; P=0.002). Greater effects were observed in patients who completed the study. Notably, the amount of weight loss was associated with the level of self-reported dietary adherence (r=0.60; P<0.001), but not with diet type (r=0.07; P=0.40). Each diet significantly reduced the ration of low-density lipoprotein cholesterol to high-density lipoprotein (HDL) cholesterol by approximately 10 percent (P<0.05), with no significant effects on blood pressure or glucose at one year. For each diet, weight loss was significantly associated with changes in the ration of total cholesterol to HDL cholesterol (r=0.36), C-reactive protein levels (r=-0.39), but again, no significant difference among the diets was observed.
The researchers concluded that although overall dietary adherence rates were low, better diet adherence was associated with greater weight loss and reductions in cardiac risk factors for each diet group. No single diet produced satisfactory adherence rates, and the progressively decreasing mean adherence scores were practically identical among the four diets. The higher discontinuation rates for the Atkins and Ornish diet groups suggest many individuals found these diets to be too extreme. In an accompanying editorial, the results of the study appear to be at odds with findings from a number of recent studies that demonstrated more weight loss, at least during the initial months, with low-carbohydrate diets. One important similarity between this trial and the other studies is a high dropout rate. Also noted the poor adherence rates found in all four diet groups, but especially the Atkins and Ornish groups.
Although patients were not allowed to choose their diet assignment, we suspected that adherence rates and clinical improvements would have been better if participants had been able to freely select from the four diet options. Our findings challenge the concept that one type of diet is best for everybody and that alternative diets can be disregarded. Likewise, our findings do not support the notion that very low-carbohydrate diets are better than standard diets, despite recent evidence to the contrary. We echoed these sentiments and stressed the importance of matching patients to a diet best suited to their needs. The right diet needs to be matched with the right patient. Ultimately, a ‘nutrigenomic’ approach most likely will be helpful. At present, there are no data to help clinicians practicably match a diet to an individual patient’s diet response genotype.