| HOME | HELP | FEEDBACK | SUBSCRIPTIONS | ARCHIVE | SEARCH | TABLE OF CONTENTS |
|
|
|
|||||||
|
|||||||||
Original Research Communications |
 |
ABSTRACT |
|---|
 TOP ABSTRACT INTRODUCTIONMETHODSRESULTSDISCUSSIONREFERENCES |
|---|
Objective: We planned to compare the relative influences of HIV infection, sex, race, and environment on body composition.
Methods: We analyzed results of body composition studies performed by bioelectrical impedance analysis in 1415 adults from 2 cohorts: white and African American men and women from the United States, and African men and women (279 HIV-infected and 1136 control). The effects of sex and HIV infection on weight, body cell mass, and fat-free mass were analyzed by using both unadjusted and age-, weight-, and height-adjusted data.
Results: Control men weighed more and had more body cell mass and fat-free mass than did control women, although control women had more fat. The strongest correlates with body composition were height and weight, followed by sex, HIV infection, age, environment, and race. Control men and women weighed more and had more body cell mass, fat-free mass, and fat than did HIV-infected men. However, differences in body composition between HIV-infected and control groups were strongly influenced by sex. Of the differences in weight between HIV-infected and uninfected subjects, fat-free mass accounted for 51% in men but only 18% in women, in whom the remainder was fat. Sex effects were similar in African and American groups.
Conclusions: Sex has a marked effect on the changes in body composition during HIV infection, with women losing disproportionately more fat than men. Sex-related differences in body composition were narrower in the HIV-infected groups. Race and environment had smaller effects than sex and HIV infection.
Key Words: Nutritional assessment • body composition • bioelectrical impedance analysis • BIA • HIV infection • malnutrition • African Americans • Africans • men • women • whites
|
INTRODUCTION |
|---|
|
TOPABSTRACTINTRODUCTIONMETHODSRESULTSDISCUSSIONREFERENCES |
|---|
There are numerous influences on body composition other than sex, including age, race, individual genetic factors, diet, physical activity, hormones, and cultural and economic factors. Although illness and injury, especially chronic disease, also affect nutritional status and body composition, the relative effects and interactions among clinical disease and these other influences on body composition have not been studied formally.
We previously compared the relative influences of sex, race, and environment on body composition in healthy adults by combining 2 databases of cohorts who had undergone bioelectrical impedance analysis (BIA): 1) HIV-seropositive and -seronegative Africans who were studied as part of Project SIDA, an international consortium of AIDS researchers from the United States, Belgium, and Zaire (12), and 2) healthy adult Americans of differing ethnic backgrounds and HIV-infected persons who have been studied at St Luke's–Roosevelt Hospital Center in New York since 1984 (13). The study design allowed simultaneous evaluation of the independent effects of sex, race (African American and white), and environment (African and African American; interpretation of the effect of environment is confounded somewhat by the genetic admixture between Africans and Europeans in African Americans, so that 
25% of the genes in an African American living in New York are European in origin; 14).
The aim of this study was to compare the relative influences of a specific disease, ie, HIV infection, and sex on body composition. The current study includes the results of studies in the control subjects reported previously (13), as well as the results of body-composition studies performed concurrently in HIV-infected African and American subjects. The data analysis was designed specifically to compare the relative effects of being male (or female) and of being HIV-seropositive (or –seronegative) after correcting for other effects on body composition, such as race, environment, height, weight, and age. The comparisons between sex and HIV infection were performed separately in American and African data sets to allow an independent replication of the results.
|
METHODS |
|---|
|
TOPABSTRACTINTRODUCTIONMETHODSRESULTSDISCUSSIONREFERENCES |
|---|
. The African groups consisted of 569 (297 men and 272 women) HIV-seronegative employees of the National Bank of Zaire (control group) and 98 (52 men and 46 women) HIV-infected persons. The American control group also consisted of 569 (202 African American and 367 white) subjects recruited by advertisement in the New York City metropolitan area; HIV testing was not performed in the American control subjects. The American HIV-infected group consisted of 115 whites and 64 African American HIV-infected individuals who have been studied at St Luke's–Roosevelt Hospital Center in New York City since 1984 (13). All subjects were between the ages 18 of and 65 y; morbidly obese subjects [body mass index (in kg/m2) >40] were excluded from analysis.
|
Statistics
All statistical analyses were performed by using SPSS (version 7.5, SPSS Inc, Chicago).
The general linear model procedure was used to test the effects of HIV status, sex, race (in the American sample), and all 2-way interactions. The analyses were run separately in the American and African samples to allow an assessment of the consistency of results as a form of cross-validation. The dependent variables were body cell mass, fat-free mass, and body fat in 3 separate analyses. Covariates were age, height, and weight. Age was transformed to improve linearity, normality of residuals, and homoscedasticity (17). The best-fit exponents for age derived by using a log-log regression were age0.056 for body cell mass, age0.141 for fat-free mass, and age0.077 for fat mass. Because the sex x HIV status interactions were significant in all models (see Results section), effect sizes were estimated after adjusting the results for age, height, and weight, and for the interaction effects of 1) sex in HIV-negative subjects, 2) sex in HIV-infected subjects, 3) HIV infection in men, and 4) HIV infection in women.
|
RESULTS |
|---|
|
TOPABSTRACTINTRODUCTIONMETHODSRESULTSDISCUSSIONREFERENCES |
|---|
). For both sexes, African Americans and whites were significantly taller than their African counterparts but similar to each other. For both sexes, African Americans were significantly heavier than Africans or whites.
Body composition in control subjects
Body cell mass and fat-free mass were higher in men than in women (P < 0.001 for all comparisons). In contrast, fat contents, both as fat mass and as a percentage of weight, were higher in women than in men (P < 0.001; Table 1). For both sexes, African Americans had more body cell mass, fat-free mass, and fat mass than did Africans (P < 0.001 for all comparisons; Table 2). In contrast, there were no significant differences in body cell mass and fat-free mass between African Americans and whites. Fat mass was significantly higher in African American women than in white women (P < 0.001; Table 2), but there were no significant differences between African American and white men. Average body cell mass and fat-free mass decreased slightly with advancing age in all subgroups, whereas fat mass rose.
|
). All HIV-infected groups had lower values for body cell mass, fat-free mass, and fat mass than the corresponding control groups. Body cell mass and fat-free mass were significantly higher in HIV-infected men than in HIV-infected women, whereas fat mass and percentage fat remained higher in the HIV-infected women than in the HIV-infected men. HIV-infected men and women had obvious sex-related differences in body composition compared with the control subjects. Of the differences in weight between HIV-infected and control women, 70–88% were due to differences in fat mass, with the rest due to lean mass. In contrast, 44–46% of the weight differences between HIV-infected and control men were due to differences in fat mass.
Multiple regression analyses were performed to identify significant determinants of body composition in disease. For each analysis, the dependent variables were body-composition results, body cell mass, fat-free mass, or fat mass, and the predictor variables were race, sex, environment, age, disease, weight, and height. The multivariate models for body cell mass, fat-free mass, and fat mass all were highly significant, with R2 ranging from 0.86 to 0.92 when all predictor variables were used (data not shown). Height and weight had the strongest influences on body cell mass, fat-free mass, and fat mass, and regression equations using these variables produced R2 values only ranging from 0.71 to 0.78. Of the other variables, sex had the strongest effect, followed by disease (HIV infection) and race. Neither age nor environment was a significant predictor in these models.
Interactions among race, sex, and HIV infection were determined separately in the American and African samples (interactions with race were not possible in the African samples). The interaction effects for race x HIV status were not significant for any of the dependent variables in the American sample. The interaction effect for race x sex was not significant for body cell mass in the American sample but was significant for fat-free mass and fat mass (P < 0.001). However, their contributions to the models in terms of R2 increment were minimal (<0.002), so these interaction terms were excluded from the models. In contrast, the interaction effects for HIV status x sex were significant for all dependent variables in both the American and African samples (P < 0.001). Therefore, the effects of HIV status must be considered separately for each sex. Similarly, the effect of sex must be considered separately for HIV-infected and -uninfected subjects. To estimate such effects, ie, HIV status effects by sex and sex effects by HIV status, contrasts of these effects were constructed and are listed in Table 3 and portrayed graphically in Figure 1. The results are shown both as the mean (±SE) of the height-, weight-, and age-adjusted results (Figure 1), and as the differences in mean values in the subgroups (Table 3).
|
|
11% of average body cell mass in control subjects compared with a difference of 2.6 kg in Africans, or 9% of average body cell mass in the control subjects. Thus, the differences in adjusted body cell mass between men and women are narrower in HIV-infected American and African subjects than in control subjects because of a larger loss of body cell mass in men.
Adjusted fat-free mass was significantly higher in men than in women, with effect sizes of 12.0 and 9.1 kg in the control American and African men, respectively, and 6.5 and 4.7 kg in the HIV-infected groups (P < 0.001 for all comparisons). Adjusted fat-free mass was also significantly lower in HIV-infected American or African men than in the respective control groups (effect sizes: 3.8 and 2.9 kg, respectively; P < 0.001 for both). The effect sizes represent 5.3% and 5.5% of average fat-free mass in American and African control subjects, respectively. In contrast, adjusted fat-free mass was relatively higher in HIV-infected American and African women than in control women (effect sizes: 1.7 and 1.6 kg, respectively; P < 0.001 for both). Once again, the differences in fat-free mass between men and women were smaller in the American and African HIV-infected groups than in the control groups. Note that the higher values for fat-free mass in HIV-infected women are artifacts of the data adjustment, especially for weight: actual fat-free mass measures in HIV-infected American and African women were 2.2–3.5 kg lower than in the corresponding control groups (Table 1).
Adjusted fat mass was higher in women than in men in both the American and African groups (effect sizes: -12.1 and -9.1 kg in control subjects, and -6.5 and -4.6 kg in HIV-infected groups; all P < 0.001). Adjusted fat mass was higher in HIV-infected than in control men, with effect sizes of -3.8 and -2.8 kg in American and African subjects (P < 0.001 for both comparisons). Once again, adjusted fat mass was higher, whereas actual fat mass was lower (Table 1). In contrast, fat mass was lower in HIV-infected than in control women (effect sizes: 1.8 and 1.6 kg, respectively: P < 0.001 for both comparisons). These differences represent 7.7% and 7.9% of the average fat mass of control American and African women, respectively. Once again, the differences in fat mass between men and women were lower in the HIV-infected than in the control groups.
Thus, the results of studies using age-, height-, and weight-adjusted data show that HIV infection exerts different effects on body composition in men and women but similar effects in Americans and Africans. Overall, the differences in body composition between control men and women were significantly less in the HIV-infected than in the control groups.
|
DISCUSSION |
|---|
|
TOPABSTRACTINTRODUCTIONMETHODSRESULTSDISCUSSIONREFERENCES |
|---|
This study was made possible through the use of BIA in 2 different cohorts. The SEEs for predicting various body compartments, 5–9%, are higher than those reported for more sophisticated measurement techniques, such as whole-body potassium counting, dual-energy X-ray absorptiometry, isotope dilution, in vivo neutron activation, computed tomography scanning, or magnetic resonance imaging (16). However, these results are not influenced by race, sex, disease, or age, and are adequate for use in large-scale, epidemiologic investigations such as the current study. BIA is an ideal method for assessing body composition in field studies because the equipment is portable, inexpensive, the measurements are simple to standardize and perform, and the results are operator-independent.
The relative differences in body cell mass between HIV-infected and control men in this study were larger than the relative differences in weight or fat-free mass. This result is consistent with previous observations that an expansion of extracellular volume masks the loss of body cell mass by limiting the observed depletion of fat-free mass or body weight (9). Studies in HIV-infected men from other laboratories have shown that body cell mass may be depleted even in the absence of weight loss (18).
Studies from another laboratory showed that the composition of lost weight in HIV-infected men was 50% fat-free mass and 50% fat (19), similar to the findings in this study. In contrast, several studies in women have suggested early or more profound loss of body fat, as also seen in this study (9, 10, 20, and SS Raghavan et al, personal communication, 1996). Thus, a single disease process, HIV infection, has a sexual dimorphism, that is, has divergent effects on changes in body composition in the 2 sexes.
It is important to understand why body composition differs in men and women under normal circumstances to learn how HIV infection could affect body composition differently in men and women. Sex hormones play a profound role in determining normal body composition. Body composition is roughly similar in boys and girls until puberty, when girls begin to accrue fat more rapidly than boys, whereas boys undergo more rapid increases in body cell mass and skeletal muscle mass (21). Menopause and surgical castration are associated with fat loss in women (22) and body cell mass depletion in men (23), respectively. Many systemic disorders are known to induce a state of secondary hypogonadism (24). Other causes of secondary hypogonadism, such as pituitary disease, are also associated with body composition changes similar to those observed in the HIV-infected groups (25). Thus, any process that produces hypogonadism could result in disproportionate body cell mass depletion in men and disproportionate fat depletion in women and convergence toward a common body composition that reflects the absence of sex hormone effects.
Many studies of HIV-infected men documented the prevalence of subnormal serum free or total testosterone concentrations (26–31) that may increase with disease progression (29). Grinspoon et al (11) correlated low free testosterone concentrations with depletion of muscle mass and decreased exercise tolerance in HIV-infected men with wasting. They also found reduced concentrations of serum insulin-like growth factor I despite elevations in mean overnight growth hormone concentrations. Several studies found no association between serum testosterone concentrations and wasting (26–28, 30), though an association was found by Coodley et al (31).
HIV-infected women also were shown to have subnormal serum testosterone concentrations (10, ES Engelson et al, personal observations, 1995). Grinspoon et al (10) correlated both amenorrhea and subnormal serum testosterone with decreased estimated skeletal muscle mass in women with AIDS wasting syndrome. Serum estradiol concentrations were lower in their subjects with amenorrhea, but the results did not correlate with body-composition measures.
The specific interactions among body composition, sex, and disease are uncertain. It is unclear which abnormality is the cause and which is the effect. Whereas hypogonadism is known to affect body composition, simple starvation is also associated with decreased serum sex hormone concentrations (32). Decreased fertility has been documented in studies of human starvation (33), and amenorrhea has been documented in studies of anorexia nervosa as well as in highly trained athletes (34, 35). Systemic diseases promote wasting through a variety of coordinated metabolic changes that may include secondary hypogonadism (24).
More recent longitudinal studies have helped clarify the issue of cause and effect. Falutz (35) performed longitudinal studies of serum testosterone, follicle-stimulating hormone, and luteinizing hormone in a group of HIV-infected men and found some subjects with normal serum testosterone concentrations but high follicle-stimulating hormone concentrations. Follow-up of these subjects showed progression to subnormal serum testosterone concentrations. Dobs et al (36) performed a retrospective, longitudinal, nested substudy of HIV-infected subjects with and without wasting and unaccompanied by a specific disease complication. They showed that both free and bioavailable testosterone concentrations fell during the 6 mo before the wasting was detected. There was a concurrent rise in serum follicle-stimulating but not luteinizing hormone content. Thus, these results suggest that hypogonadism may precede malnutrition.
A plausible explanation for the current results as well as those in the literature is that HIV infection promotes an acquired hypogonadal state in addition to other nutritional effects mediated through changes in food intake (37, 38), nutrient absorption, and energy expenditure (37–39). Reduced testosterone should result in disproportionate depletion o body cell mass and skeletal muscle mass in men, whereas decreased estrogen and progesterone effects lead to predominant fat loss in women.
An alternative hypothesis is that body-composition changes during weight loss are related mainly to premorbid body composition, as shown by Forbes (40), as well as to the severity of the nutritional insult. According to this hypothesis, women lose more fat than men do simply because they have more fat in the premorbid state. This hypothesis could not be tested in the current study because all female control groups had higher percentage body fat than all male control groups. To directly address the question, both lean and obese men and women would have to be followed longitudinally, with changes in weight and body composition determined prospectively. The argument is circular, however, because sex hormones play a role in defining premorbid body composition, and the observed changes in body composition during disease imply an alteration in sex hormone effect, if not in hormone production and circulating concentrations.
Further studies are needed to test the hypothesis that the changes in body composition during systemic illnesses, in both HIV-related and other disease processes, are affected by altered sex hormone effects. Ideally, the studies should be longitudinal and controlled for premorbid body composition, specific etiology of disease, pathogenic mechanism underlying wasting, and severity of the energy deficit. The specificity of the body-composition change in HIV-infected patients with hypogonadism should be determined, as well as a potential biological gradient between the degree of hypogonadism and the degree of body-composition change, a temporal relation between hormonal and body-composition changes, and the reversibility of the specific body-composition changes with targeted interventions. Studies in both HIV-infected (27) and -uninfected (41, 42) persons have shown that testosterone therapy is associated with body cell mass repletion. No studies have been published investigating estrogen or progesterone replacement in secondary hypogonadism induced by systemic disease in women, though hormone replacement may mitigate the body-composition changes that occur during menopause in women (43). Studies should be performed to determine whether a sexual dimorphism exists in other states of malnutrition.
HIV infection produces different body-composition alterations in men and women. These alterations may result from disease-induced hypogonadism.
|
ACKNOWLEDGMENTS |
|---|
|
FOOTNOTES |
|---|
2 Supported by NIH grants DK 42618, DK 37352, AI 26695, the Centers for Disease Control and Prevention, and Project SIDA.
3 Address reprint requests to DP Kotler, GI Division/S&R 1301, St Luke's–Roosevelt Hospital Center, 1111 Amsterdam Avenue, New York, NY 10025. E-mail: dpkotler{at}aol.com.
|
REFERENCES |
|---|
|
TOPABSTRACTINTRODUCTIONMETHODSRESULTSDISCUSSIONREFERENCES |
|---|
This article has been cited by other articles:
|
|
 |
|
  E. Aghdassi, B. Arendt, I. E. Salit, and J. P. Allard Estimation of Body Fat Mass Using Dual-Energy X-Ray Absorptiometry, Bioelectric Impedance Analysis, and Anthropometry in HIV-Positive Male Subjects Receiving Highly Active Antiretroviral Therapy JPEN J Parenter Enteral Nutr, March 1, 2007; 31(2): 135 - 141. [Abstract] [Full Text] [PDF]
|
|
|
|
 |
|
 F. Visnegarwala, S. S Raghavan, C. M Mullin, G. Bartsch, J. Wang, D. Kotler, C. L Gibert, J. Shlay, C. Grunfeld, A. Carr, et al. Sex differences in the associations of HIV disease characteristics and body composition in antiretroviral-naive persons Am. J. Clinical Nutrition, October 1, 2005; 82(4): 850 - 856. [Abstract] [Full Text] [PDF]
|
|
|
|
|
|
P. C Papathakis, N. C Rollins, K. H Brown, M. L Bennish, and M. D Van Loan Comparison of isotope dilution with bioimpedance spectroscopy and anthropometry for assessment of body composition in asymptomatic HIV-infected and HIV-uninfected breastfeeding mothers Am. J. Clinical Nutrition, September 1, 2005; 82(3): 538 - 546. [Abstract] [Full Text] [PDF]
|
|
|
|
|
|
J. E Forrester, D. Spiegelman, E. Tchetgen, T. A Knox, and S. L Gorbach Weight loss and body-composition changes in men and women infected with HIV Am. J. Clinical Nutrition, December 1, 2002; 76(6): 1428 - 1434. [Abstract] [Full Text] [PDF]
|
|
|
|
 |
|
 H. Friis, E. Gomo, N. Nyazema, P. Ndhlovu, P. Kaestel, H. Krarup, and K. F. Michaelsen HIV-1 Viral Load and Elevated Serum {alpha}1-Antichymotrypsin Are Independent Predictors of Body Composition in Pregnant Zimbabwean Women J. Nutr., December 1, 2002; 132(12): 3747 - 3753. [Abstract] [Full Text] [PDF]
|
|
|
|
|
|
E. Villamor, G. Msamanga, D. Spiegelman, G. Antelman, K. E Peterson, D. J Hunter, and W. W Fawzi Effect of multivitamin and vitamin A supplements on weight gain during pregnancy among HIV-1-infected women Am. J. Clinical Nutrition, November 1, 2002; 76(5): 1082 - 1090. [Abstract] [Full Text] [PDF]
|
|
|
|
 |
|
 P. Stenvinkel, P. Barany, S. H. Chung, B. Lindholm, and O. Heimburger A comparative analysis of nutritional parameters as predictors of outcome in male and female ESRD patients Nephrol. Dial. Transplant., July 1, 2002; 17(7): 1266 - 1274. [Abstract] [Full Text] [PDF]
|
|
|
|
|
|
S. Shah, C. Whalen, D. P. Kotler, H. Mayanja, A. Namale, G. Melikian, R. Mugerwa, and R. D. Semba Severity of Human Immunodeficiency Virus Infection Is Associated with Decreased Phase Angle, Fat Mass and Body Cell Mass in Adults with Pulmonary Tuberculosis Infection in Uganda J. Nutr., November 1, 2001; 131(11): 2843 - 2847. [Abstract] [Full Text] [PDF]
|
|
|
|
 |
|
 D. P. Kotler Cachexia Ann Intern Med, October 17, 2000; 133(8): 622 - 634. [Abstract] [Full Text] [PDF]
|
|
|
|
|
|
C. Corcoran, E. J Anderson, B. Burrows, T. Stanley, M. Walsh, A. M Poulos, and S. Grinspoon Comparison of total body potassium with other techniques for measuring lean body mass in men and women with AIDS wasting Am. J. Clinical Nutrition, October 1, 2000; 72(4): 1053 - 1058. [Abstract] [Full Text] [PDF]
|
|
|
|
 |
|
 D. P. KOTLER Body Composition Studies in HIV-Infected Individuals Ann. N.Y. Acad. Sci., May 1, 2000; 904(1): 546 - 552. [Abstract] [Full Text] [PDF]
|
|
|
|
 |
|
 K. J. Ellis Human Body Composition: In Vivo Methods Physiol Rev, April 1, 2000; 80(2): 649 - 680. [Abstract] [Full Text] [PDF]
|
|
| ||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||
| HOME | HELP | FEEDBACK | SUBSCRIPTIONS | ARCHIVE | SEARCH | TABLE OF CONTENTS |
) and women (
). Separate analyses were performed in American (A, C, and E) and African (B, D, and F) groups. The differences in body cell mass, fat-free mass, and fat mass between men and women were much larger in the control than in the HIV-infected groups. Note the different scales in the American and African samples.