点击显示 收起
Department of Physical Education Sports Science and Recreation Management Loughborough University Leicestershire LE11 3TU United Kingdom E-mail d.j.stensel{at}lboro.ac.uk
Dear Sir:
The cross-sectional study of Polito et al (1) examined the relation between basal metabolic rate (BMR) and anorexia nervosa. In the introduction to their article, Polito et al state that many studies showed a reduced BMR in subjects with anorexia nervosa and that "the main controversy remains whether the decrease in BMR is due to a change in body composition or whether it represents a down-regulation of cellular metabolism." These authors conclude from their findings that BMR is depressed in subjects with anorexia nervosa and that this is not explained by changes in body composition. This suggests that "... the metabolic activity of the active tissue mass may have been reduced" in subjects with anorexia nervosa.
I am not sure that the conclusions of Polito et al are justified because these authors did not adequately control for differences in body composition. Although they did control appropriately for fat-free mass by using analysis of covariance, their data (Table 1 of the article) show that the biggest difference in body composition between subjects with anorexia nervosa and control subjects is in fat mass and not fat-free mass. Fat mass was 57% lower in patients with anorexia nervosa than in control subjects (6.8 compared with 15.7 kg, respectively), whereas fat-free mass was only 16% lower (34.7 compared with 41.2 kg). It is now recognized that fat mass also makes a contribution to BMR (2) and recent studies controlled for both fat mass and fat-free mass when comparing BMR between groups (3–6).
If Polito et al had used fat mass and fat-free mass as covariates when comparing BMR, it is possible that group differences would have been eliminated, which would suggest that body composition differences were responsible for the reduced BMR values in subjects with anorexia nervosa. Such a contention is supported by the fact that when Polito et al used body weight as a covariate instead of fat-free mass they found no difference in BMR between subjects with anorexia nervosa and control subjects. In fact, if a single covariate is to be used for the data of Polito et al, it would be more appropriate to use body weight than to use fat-free mass. The rationale for using fat-free mass as a single covariate is that it is the best predictor of BMR (7). However, the data of Polito et al show that, for their subjects, body weight and not fat-free mass was the best BMR predictor (r2 = 0.62 and 0.48 for body weight and fat-free mass, respectively; Figure 1 of the article).
Thus, a plausible explanation for the lower BMRs in subjects with anorexia nervosa after control for fat-free mass is their greatly reduced fat mass. To eliminate this possibility, Polito et al could reanalyze their data using analysis of covariance to control for both fat mass and fat-free mass. This would either strengthen their conclusion that reductions in BMR with anorexia nervosa are not due to body composition changes or invalidate that conclusion.
REFERENCES