| HOME | HELP | FEEDBACK | SUBSCRIPTIONS | ARCHIVE | SEARCH | TABLE OF CONTENTS |
|
|
|
|||||||
|
|||||||||
| ||||||||||||||||||||||||||||||||||||||||||||||||||||||||||
Letters to the Editor |
Department of Nutrition Harvard School of Public Health 665 Huntington Avenue Boston, MA 02115
Dear Sir:
Purnell et al have interpreted the fine study by Knopp et al (1) strangely and have ignored the very reason for conducting randomized trials—to account for factors other than the specific intervention that might change over time and thus influence the outcome being evaluated. The reason for conducting such a randomized trial is to compare the change in the treatment group with the change in the parallel, control group and not to evaluate the significance of changes within groups. Important strengths of their study were that a large number of subjects were randomly assigned to receive 1 of 4 fat intakes and that each group received dietary counseling and monitoring. This allowed valid comparisons of the effects of different fat intakes on changes in weight. The fact that all groups lost 
3 kg indicates strongly that the percentage of energy from fat, over the range studied, did not influence body weight and that some other factor common to all 4 groups did. Purnell et al stated that this common factor was the reduction in fat intake in all groups compared with baseline, even though weight reduction was not related to the decrease in fat intake. Although the common factor cannot be proven directly from this study, the Hawthorne effect is a more likely explanation. Specifically, that raising ones consciousness about food intake combined with intensive support, counseling, and feedback will result in a modest weight loss regardless of the percentage of energy from fat in the diet. Indeed, dietitians have long known that careful recording and monitoring of dietary intakes is an important component of weight control.
The finding of an effect of intervention unrelated to fat intake by Knopp et al highlights the methodologic shortcomings of most randomized trials of fat reduction and emphasizes the desirability of including a control group with a similar intensity of intervention in such studies. For example, interventions aimed at reducing total energy or carbohydrate intakes would be appropriate comparison strategies. The other studies cited by Purnell et al lacked such a control group. For example, in the trial by Sheppard et al (2), one group was given intensive instruction in fat reduction followed by careful monitoring via weighed food records, but the control group received no such intervention; the differences in the changes in weight between the 2 groups were -2.6 kg at 1 y and -1.8 kg at 2 y. The small number of studies on the effects of fat reduction on body weight in which the control group did receive intervention comparable with that of the treatment group showed minimal or no effects on body weight: none in Knopp et al's study (1) when properly analyzed, a loss of 1.4 kg in the study by Jeffery et al (3), and a loss of 0.8 kg in the National Diet Heart Study (4). Thus, the best evidence from long-term studies involving control groups with a similar level of intervention intensity show extremely little if any effect of the percentage of energy from fat on body weight. However, even if the more optimistic but biased assumption of a 2–3-kg effect were accurate, this is still an imperceptible and clinically unimportant change for an overweight or obese person. If the proponents of low-fat diets were more candid about the weight loss expected from a major change in diet, there would likely be few "takers," but less disillusionment and loss of credibility for the nutrition community. Patients prescribed a low-fat diet should be informed that such a diet is likely to increase serum triacylglycerol and reduce HDL-cholesterol concentrations (5), which are associated with a higher risk of coronary artery disease.
Purnell et al cite the findings of Schaefer et al's study (6) to suggest that changes in weight may vary in persons consuming low-fat diets, possibly because of genetic differences. However, this study lasted only 12 wk and longer-term evidence suggests that these weight changes would not be maintained. Variations in weight changes in response to changes in the mix of macronutrients are indeed possible, but genetic modifiers are yet to be identified. Because the average effect on body weight of a reduction in fat intake is so small in the longer-term studies conducted to date, if there is a common subgroup that responds with a major weight reduction, we would also need to hypothesize another subgroup that responds with weight gain.
Anyone would agree with Purnell et al that regular exercise will benefit most patients, but available evidence strongly suggests that a focus on energy intakes from fat with no regard to energy intake from carbohydrate will have little effect on body weight. Moreover, there is overwhelming evidence from prospective studies (7) and randomized trials (8) that replacement of saturated and trans fats with unsaturated fat in the diet will substantially reduce the risk of coronary artery disease, but that replacement of fat with carbohydrate in the diet will have little if any effect (9). The misguided focus on a reduction of dietary fat per se to reduce body weight has resulted in a lost opportunity to have a major effect on the most important cause of death in Western countries.
REFERENCES
| ||||||||||||||||||||||||||||||||||||||||||||||||||||||||||
| HOME | HELP | FEEDBACK | SUBSCRIPTIONS | ARCHIVE | SEARCH | TABLE OF CONTENTS |
