| HOME | HELP | FEEDBACK | SUBSCRIPTIONS | ARCHIVE | SEARCH | TABLE OF CONTENTS |
|
|
|
|||||||
|
|||||||||
Letter to the Editor |
1 Department of Nutrition Sciences University of Alabama at Birmingham Birmingham, AL 25294
2 Department of Nutrition Sciences Clinical Nutrition Research Center University of Alabama at Birmingham Birmingham, AL 25294 E-mail: weinsier{at}shrp.uab.edu
3 Department of Human Studies Clinical Nutrition Research Center University of Alabama at Birmingham Birmingham, AL 25294
Dear Sir:
We wish to respond to many issues that Pereira et al raised regarding the methods used in our recently published study (1). First, Pereira et al described our study sample as consisting of "2 very different groups of women" and suggested that a better study design would be one in which the subjects were derived from the same population. The reason provided for this assertion was that there may be important behavioral or metabolic differences between persons with and without a recent history of weight loss.
Our intent in selecting both previously overweight and previously normal-weight subjects was to obtain a baseline study group that would give a wide range of weight-gain responses during follow-up. In fact, as was outlined in our article, the subjects were phenotypically homogeneous, ie, percentage body fat; metabolic characteristics of resting energy expenditure, fuel utilization, or insulin sensitivity; activity energy expenditure (AEE); and physical activity (activity-related time equivalent index) were not significantly different between the 2 groups of women. Consequently, we viewed our subjects as one group that was likely to vary widely in the propensity toward obesity on the basis of their weight histories. For example, some of the previously normal-weight women were likely to gain weight over time, simply because they were still young and had not yet had a chance to gain much weight; 14% of the 28 women recruited with no history of overweight did not maintain their baseline weight status over the year. By contrast, whereas some of the weight-reduced women were likely to gain a significant amount of weight, this was not the case for all of them. This was evidenced by the fact that, whereas 95% of the "gainers" were previously overweight women, only 58% of the previously overweight women were gainers. Thus, although the subjects should have provided a range of weight-gain responses enabling investigation of factors relating to weight gain and weight maintenance, they could not be considered 2 distinct groups in terms of an a priori assumption that they would or would not gain weight.
We agree with the rationale that there may be important behavioral or metabolic differences between having or not having a recent history of weight loss, and our data support this notion. We did not report dietary intake patterns in this article, and we are not convinced that such reports would be as accurate or as objective as were our measures of sleeping energy expenditure (SEE) and AEE, which were measured by chamber calorimetry and the doubly labeled water method, respectively. Our study showed no significant differences in patterns of SEE, fuel utilization, or exercise economy between the "maintainers" and the gainers. Furthermore, because we had reasonably good measures of body composition, we were able to calculate the energy cost associated with accumulating the weight gained by the gainers over the year. The difference in energy expended in physical activity between the gainers and maintainers accounted for 77% of the energy cost of weight gain. Given that SEE did not change, we concluded that physical activity has a strong influence on variations in relative weight and adiposity. If our calculations are correct, then AEE appeared to play the major role in the weight gain and would have been more important in this group of subjects than even dietary intake. Luke et al (2) recently confirmed that AEE is strongly associated with adiposity and, with age, sex, and weight, explained 77% of the variance in percentage body fat. However, despite the importance of AEE, we certainly do not discount the role that dietary intake plays in altered weight status.
Finally, we agree with the suggestion by Pereira et al that evaluating the nature of differences in physical activity (intensity, duration, and frequency) would be valuable in better understanding the role that physical activity and exercise play in preventing weight gain. To this end, we have undertaken a study to evaluate the influence of aerobic and resistance training on weight maintenance after weight loss and the nature of self-selected, free-living physical activity displayed by gainers and maintainers. Unfortunately, we did not have good descriptors for the nature of the activity patterns of the subjects in our completed study (1).
Finally, a point raised by Pereira et al regarding the interpretation of our statistical analyses needs clarification. In repeated-measures analyses of longitudinal data, the analyses are not reduced to a cross-sectional comparison if the measured variable does not change over time. In fact, cross-sectional analyses do not lend themselves to deductions about cause and effect relations. In our study, wherein the gainers had lower levels of physical activity and fitness than did the maintainers, in both their normal- and weight-gain states, the data indicate clearly that weight gain per se was not the cause of the lower physical activities or fitness levels. Conversely, the data suggest that lower activities and fitness levels may well have contributed to the weight gain observed in our subjects.
REFERENCES
| ||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||
| HOME | HELP | FEEDBACK | SUBSCRIPTIONS | ARCHIVE | SEARCH | TABLE OF CONTENTS |
