点击显示 收起
1 From the Department of Biostatistics, University of Alabama at Birmingham.
2 Reprints not available. Address correspondence to DB Allison, Department of Biostatistics, Ryals Public Health Building, Suite 327, The University of Alabama at Birmingham, 1665 University Boulevard, Birmingham, AL 35294-0022. E-mail: dallison{at}uab.edu.
See corresponding article on page1368.
Time and again, as we study physiologic phenomena and their anatomic correlates, greater clarity comes from realizing that heterogeneous wholes are composed of somewhat more homogeneous parts and from examining the independent and interactive effects or associations of these parts with outcomes of interest. For example, whereas previously we analyzed total cholesterol and its association with health, few scientists today would not measure separately the HDL and LDL cholesterol (as well as other components and subfractions), given their opposite effects on cardiovascular disease risk. Similarly, recent authors (1) teach us that not all fatty acids can be lumped in a single category of fat when we consider the effects of dietary fat intake. More than 150 y ago, when Quetelet developed the body mass index, expressed as weight divided by the square of stature, one might have wondered why he did not divide by the cube of stature. Indeed, because people are 3-dimensional beings, one might expect weight to increase in proportion to the cube of stature. This would be so if people were spheres of uniform density. However, it so happens that, among adults (who are not spheres of uniform density), weight increases approximately in proportion to the square of stature. Quetelet’s early observation foreshadowed the importance of considering tissue heterogeneity in obesity research. More than half a century ago, Vague (2) helped us realize that, by differentiating among anatomical depots of fat, we could better appreciate the health effects of adiposity. That fecund observation has been a driving force in our field ever since.
Only in the past few years have obesity researchers begun to differentiate between components of soft-tissue lean body mass (LBM) (3–5). In this issue of the Journal, Byrne et al (6) help point the way toward a greater understanding of certain commonly observed ethnic differences by documenting and evaluating the effect of differential changes in the mass of various components of lean body tissue.
African American women are well documented to have a greater prevalence of obesity than do European American women. African American women also are, on average, less successful at losing weight and maintaining a reduced weight through conventionally offered weight-loss programs (7). The reasons for these ethnic differences are not currently understood. Several years ago, a flurry of reports showed that, even after control for LBM, African Americans had lower resting energy expenditure than did European Americans (7). Now Byrne et al (6) further subfractionate LBM to give us additional insight. They found that, when adjusted for total LBM and fat mass, resting energy expenditure was lower in African American women than in European American women both before and after weight loss. However, these ethnic differences were not observed when adjusted for regional LBM, ie, trunk LBM and limb LBM. Presumably, in terms of soft tissue, LBM on the trunk is composed of a high proportion of LBM from organs, whereas LBM on the limbs is composed solely of skeletal muscle. Although the authors’ measurement system did not allow them to measure organ volumes per se, it seems a fairly obvious and reasonably safe conjecture that the differences observed were due to differences in organ volume. However, there are other possibilities, including the fact that organs of any fixed volume may have different metabolic rates in different ethnic groups.
Of course there are methodologic challenges and interpretive issues in such a study. In using body mass index and family history of obesity as selection criteria, the authors may have selected for subjects from different portions of their ethnic-specific distributions of obesity and genetic propensity for obesity. That is, the sample of European American subjects selected was fatter and more predisposed to obesity relative to the overall white population than was the sample of African American subjects relative to the overall black population. Although this effectively controlled for current body mass index, it may have introduced different biases.
Studies of energy restriction in rodents show that, with the exception of brain and testes, organs tend to achieve masses that are roughly proportional to total body mass (8). To the extent that weight loss and regain in humans alter this expected proportionality, might this weight loss and regain have differential effects? Earlier research has shown that weight loss may have differential effects on longevity depending on the proportion of the weight lost as fat or lean mass (9). If future research goes further by defining the subcomponents of changes in lean mass, we may gain still further insights. Previous research also suggested that we might promote greater weight maintenance through resistance exercise than through aerobic exercise, under the assumption that the former would build LBM, thereby increasing the metabolic rate and making it easier to maintain a reduced weight. However, results have been minimally supportive at best (10). In hindsight, this might be expected, because much of the metabolic activity of LBM is not in the skeletal muscle but in the organs. Refocusing the field of body composition and energetics from simply studying undifferentiated lean mass to studying lean mass and its components may help us better understand these issues.
To the extent that preservation of LBM in general or of particular subcomponents of LBM is important to achieving health benefits with weight loss or to making weight loss easier to maintain, future research may investigate combined treatment approaches in which one treatment component produces an energy deficit and another affects the anatomic depots from which body energy is drawn. For example, recent work showing that antibodies to myostatin can increase muscle mass in certain mice (11) suggests the possibility of administering such agents to persons undergoing weight loss. Members of some ethnic groups might benefit more from such treatment than would members of other ethnic groups.
Another interesting result of the study by Byrne et al is that, after weight regain, the women in both groups regained their limb LBM but did not fully regain their trunk LBM. Such a result might cause a reduced metabolic rate after weight regain and favor further weight regain. This and related questions are controversial, and data on whether metabolic rates, after control for total LBM, are altered after weight loss or regain are equivocal at best. However, if such differences were present, these changes in the distribution of LBM might offer a potential explanation. However, for those who see evidence of such metabolic alterations in response to weight loss, weight regain, or both as evidence of a set point, such an explanation might be disappointing because it invokes a passive explanation (ie, a settling point) rather than an active, "smart" system (ie, a set point).
Byrne et al ended their article by noting that "these results show that, in comparing energy expenditure between race, adjustment for differences in distribution of LBM may have to be considered." I agree with their point, but I believe the implications can be extended further. Most weight-loss studies still examine only body weight, rather than body-composition. The results reported by Byrne et al, as well as those reported in the continually expanding field of body-composition research, suggest that body-composition measurements are an important part of understanding the benefits, effects, and patterns of weight loss. Byrne et al go further by showing that we should consider body composition not only in terms of percentage body fat and body fat distribution but also in terms of distribution of LBM.
REFERENCES
Related articles in AJCN: