点击显示 收起
1 From the Department of Nutrition and Health, Research Institute of Child Nutrition, Dortmund, Germany (TD and TR), and the Steroid Laboratory, Department of Pharmacology, University of Heidelberg, Heidelberg, Germany (CM-G).
2 Supported by the Ministerium für Wissenschaft und Forschung des Landes Nordrhein-Westfalen and by the Bundesministerium für Gesundheit. 3 Address reprint requests to T Remer, Forschungsinstitut für Kinderernährung (Research Institute of Child Nutrition), Heinstück 11, 44225 Dortmund, Germany. E-mail: remer{at}fke-do.de.
ABSTRACT
Background: Excess endogenous or exogenous cortisol is a potent stimulus for fat gain.
Objective: We examined whether physiologic variations in endogenous cortisol secretion may be associated with changes in body composition during growth.
Design: Anthropometric measurements and 24-h excretion rates of urinary free cortisol (UFF) and cortisone (UFE) and the sum of 3 major glucocorticoid metabolites (GC), which reflects overall daily cortisol secretion, were determined cross-sectionally in healthy preschool (50 boys and 50 girls aged 4–5 y), late prepubertal (50 boys and 50 girls aged 8–9 y), and pubertal (50 males aged 13–14 y and 50 females aged 12–13 y) subjects.
Results: Significant positive associations (P < 0.001) were found between GC excretion and fat mass, percentage body fat, and body mass index by using covariance analysis adjusted for the grouping factors sex and age. The relations between GC and indexes of body fat remained significant (P < 0.05) even after GC was corrected for individual body surface area and the effect of maternal body mass index on fatness was considered. No consistent associations with fat indexes were seen for UFF, UFE, or the ratio of major urinary cortisol to cortisone metabolites, which reflects 11ß-hydroxysteroid dehydrogenase type 1 activity.
Conclusions: Although direct effects of UFF and UFE on body composition were not shown, our findings strongly suggest that a higher adrenocortical activity is one endocrine-metabolic feature of healthy children with higher body fat. Whether urinary GC is a long-term predictor of fat gain during childhood should be analyzed in future studies.
Key Words: Adolescents • adrenocortical activity • body mass index • BMI • body composition • body fat • children • cortisol • cortisone • glucocorticoid metabolites • 11ß-hydroxysteroid dehydrogenase • 24-h urine collection
INTRODUCTION
Most of the genetically determined potential predictors of weight gain are endocrine-metabolic factors. An endocrine-metabolic predictor of fat gain that has been known for a long time is endogenous and exogenous hypercortisolism. In both clinical and experimental settings, cortisol excess is associated with profound alteration of intermediary metabolism, characteristically resulting in obesity, insulin resistance, and changes in lipid metabolism (1, 2). In vitro, glucocorticoids promote preadipocyte differentiation and favor cellular lipid accumulation in an anabolic state with appropriate insulin availability (3–5). Recent studies in healthy adults indicate that even a moderately increased cortisol secretion may contribute to increased body fat (6, 7).
Because childhood obesity is increasing rapidly throughout the world (8), it is important to clarify whether endogenous glucocorticoids may also be involved in fat gain during growth. Until now, this relation was less well examined in children and adolescents. Only a few previous studies assessed the association between serum cortisol concentrations (9, 10) or urinary glucocorticoid excretion (11, 12) and indexes of body fatness in children. However, none of these studies examined the cortisol secretion status together with the activity of 11ß-hydroxysteroid dehydrogenase type 1 (11ß-HSD1; EC 1.1.1.146). This enzyme, which is partly expressed in adipose tissue, regenerates active cortisol from inactive cortisone. Current debate suggests that increased adipocyte 11ß-HSD1 may be involved in fat gain in humans (13). Therefore, we reexamined the relation between cortisol status and body composition by measuring variables of body fatness and daily cortisol secretion as well as 11ß-HSD1 activity in a large sample of healthy children and adolescents. Because maternal body mass index (BMI) was shown to affect the fatness of children (14, 15), its influence was also considered.
SUBJECTS AND METHODS
Subjects, study design, and anthropometry
The study was performed in 300 healthy children and adolescents (150 males and 150 females) who were all participants in the DONALD (Dortmund Nutritional and Anthropometric Longitudinally Designed) study, an ongoing observational study investigating the interrelations among nutrition, metabolism, and development during growth (16). At least once a year, the subjects were medically examined and anthropometric measurements were obtained. Generally, a 24-h urine sample was collected in yearly intervals on the last day that the subjects kept a 3-d weighed diet record. Each time, the subjects were carefully instructed in the collection procedure for the 24-h urine sample. The subjects were healthy, participating in their normal activities, and consuming their usual diets on the day of the urine collection (17). During the 24-h collection period, all micturitions were stored immediately in preservative-free, Extran-cleaned (Extran MA03; Merck, Darmstadt, Germany) 1-L plastic containers at temperatures < -12 °C (in freezers at home) before transfer to the research institute (where the storage temperature was -20 °C until thawing for urine analyses). The study was approved by the institutional review board of the Research Institute of Child Nutrition Dortmund.
The children were selected according to pubertal stages and specific age groups as follows: group 1 (preschool children aged 4–5 y, 50 boys, 50 girls; Tanner stage 1), group 2 (late prepubertal children aged 8–9 y, 50 boys, 50 girls; Tanner stage 1), and group 3 (50 pubertal males aged 13–14 y and 50 pubertal females aged 12–13 y; both sexes: Tanner stages 3–5). The stages of pubertal development were determined by a physician with the use of the grading system defined by Tanner for pubic hair.
Body weight was measured to the nearest 0.1 kg with an electronic scale (Seca 753 E; Seca Weighing and Measuring Systems, Hamburg, Germany) and body height to the nearest 0.1 cm with a digital telescopic wall-mounted stadiometer (Harpenden, Crymych, United Kingdom). Triceps, biceps, subscapular, and suprailiac measurements were made on the right side of the body with a Holtain skinfold caliper (Holtain Ltd, Crosswell, United Kingdom). Two equations were used to calculate body density from 4 skinfold-thickness measurements for the 3 age groups. The formula reported by Brook (18), which is more appropriate for prepubertal children and can be used from the age of one, was applied for group 1. For groups 2 and 3, the formulas of Deurenberg et al (19) were applied; these—in contrast to other skinfold-thickness equations—showed no mean bias for fatness when checked against a 4-component model of body-composition analysis (20). Percentage body fat (%BF) was then derived by using the equation of Siri (18, 19). Fat mass (FM; in kg) was obtained by multiplying %BF by body weight. Fat-free mass was calculated by subtracting FM from body weight.
Hormone measurements
Urinary free cortisol (UFF), urinary free cortisone (UFE), tetrahydrocortisone (THE), tetrahydrocortisol (THF), and 5-THF were measured by specific radioimmunoassays with the use of tritiated steroids (Amersham Pharmacia Biotech, Freiburg, Germany) and specific antibodies, raised and characterized in our laboratory, as described elsewhere (21). Before radioimmunoassay, UFF and UFE were extracted from the urine with dichloromethane and chromatographically purified by using Celite columns (Celite columns 545 AW; Sigma-Aldrich Chemie GmbH, Steinheim, Germany). THE, THF, and 5-THF were quantified after treatment with ß-glucuronidase (Roche Diagnostics GmbH, Mannheim, Germany) in a final dilution of 1:1200 (vol:vol). Intra- and interassay CVs were < 10% and < 13%, respectively. The sum of the 3 major urinary glucocorticoid metabolites, THE, THF, and 5-THF, which reflects overall daily cortisol secretion, is subsequently denoted as GC.
Assessment of 11ß-hydroxysteroid dehydrogenase activities
Corticosteroid action is, in part, regulated at a prereceptor level by 11ß-HSD, which catalyzes the reversible interconversion of cortisol with its inactive metabolite cortisone (22–24). Two isoforms of this enzyme have been characterized: 11ß-HSD1 usually catalyzes the reduction (reactivation) of cortisone to cortisol and is expressed in many tissues, including liver and subcutaneous and visceral fat. By contrast, 11ß-HSD2 inactivates cortisol to cortisone and thus protects the nonselective mineralocorticoid receptor in the kidney and colon from cortisol excess (22–24). Assessment of 11ß-HSD activities in vivo was done conventionally by calculating the urinary ratios of either steroid ring A–reduced metabolites of cortisol and cortisone [(5-THF + THF)/THE = 11ß-HSD1] (22, 23) or free cortisol and free cortisone (UFF/UFE = 11ß-HSD2) (22, 23).
Statistical analysis
Results are expressed as means ± SDs. Comparisons of data among the 3 groups and sexes were initially performed by two-way analysis of variance. Because in this analysis significant sex-by-age group interactions were found for some variables (eg, for %BF or 11ß-HSD1), the analyses of the different variables were performed for each sex independently by using a one-way analysis of variance. The variable means of the age groups were compared by using the Tukey test.
Simple regression analysis was undertaken to examine the relation between unadjusted daily GC excretion and indexes of body fatness (fat mass, BMI) in each sex separately. This was followed by an analysis of covariance (ANCOVA) with the use of the general linear model procedure to determine whether the relation between GC and indexes of fatness was significant after adjusting for sex and age group. In a separate ANCOVA, we examined whether there was still a relevant glucocorticoid effect after GC excretion was corrected for individual body surface area and after another potential determinant of body composition, the mother’s BMI, was included. All tests were two-tailed. P < 0.05 was considered statistically significant. Analyses were performed with SAS for WINDOWS (version 6.12; SAS Institute, Inc, Cary, NC).
RESULTS
The physical characteristics of the children are shown in Table 1. As expected, BMI and FM increased from preschool to puberty in both sexes. %BF increased during prepuberty in girls, whereas in boys it dropped (accordingly, there was a significant %BF-by-sex interaction; Table 1). Excretion rates of UFF, UFE, and the sum of THE + THF + -THF (GC) rose strongly with age in boys and girls (Table 1). When all 300 children were considered as a group, there were no significant sex differences in UFF (boys: 11.70 ± 5.24 µg · m-2 · d-1; girls: 10.61 ± 5.34 µg · m-2 · d-1; P = 0.075) and UFE (boys: 21.02 ± 7.62 µg · m-2 · d-1; girls: 20.97 ± 8.88 µg · m-2 · d-1; P = 0.956) after normalization to individual body surface area (BSA). However, BSA-adjusted GC excretion was significantly higher in boys than in girls (boys: 4.14 ± 1.38 mg · m-2 · d-1; girls: 3.78 ± 1.19 mg · m-2 · d-1; P = 0.0167). For 11ß-HSD1, significant changes were observed with age in girls (decrease into puberty) and boys (decrease only before puberty) (Table 1). Variations were not significant for renal 11ß-HSD2.
View this table:
TABLE 1 . Basic characteristics, urinary excretion of glucocorticoid metabolites, and activities of 11ß-hydroxysteroid dehydrogenase type 1 (11ß-HSD1) and 11ß-HSD2 in 300 healthy children and adolescents according to age group1
No relation was seen between UFF, UFE, 11ß-HSD1, or 11ß-HSD2 and body composition (expressed as FM and BMI) in any sex and age group. With regard to %BF, only in the subgroup of pubertal boys was a significant, negative association seen with 11ß-HSD1 (r = -0.41, P < 0.01). However, positive relations were found between GC excretion and FM for all age groups in girls and for group 2 in boys (Figure 1). Similar associations occurred between GC excretion and %BF. Also, BMI was positively associated with GC in all groups in girls and in groups 2 and 3 in boys (Figure 2).
FIGURE 1. . Associations of urinary 24-h excretion rate of GC (the sum of the 3 major glucocorticoid metabolites, tetrahydrocortisone + tetrahydrocortisol + 5-tetrahydrocortisol) with fat mass in 3 age groups of healthy children [n = 300, 50 boys (•) and 50 girls ()/group]. R2 denotes the regression coefficient from a simple regression analysis of unadjusted daily GC excretion. A subsequent analysis of covariance, with sex and age group as grouping factors and daily GC excretion as covariate, yielded the following P values for the associations with body fat: sex, P = 0.223; age group, P < 0.001; GC, P < 0.001; GC-by-sex interaction, P = 0.001; GC–by–age group interaction, P = 0.096.
FIGURE 2. . Associations of urinary 24-h excretion rate of GC (the sum of the 3 major glucocorticoid metabolites, tetrahydrocortisone + tetrahydrocortisol + 5-tetrahydrocortisol) with BMI in 3 age groups of healthy children [n = 300, 50 boys (•) and 50 girls ()/group]. R2 denotes the regression coefficient from a simple regression analysis of unadjusted daily GC excretion. A subsequent analysis of covariance, with sex and age group as grouping factors and daily GC excretion as covariate, yielded the following P values for the associations with BMI: sex, P = 0.284; age group, P = 0.004; GC, P < 0.001; GC-by-sex interaction, P = 0.238; GC–by–age group interaction, P = 0.108.
Covariance analysis, with absolute daily GC excretion as covariate, confirmed the significant association of GC with indexes of fatness after the grouping factors sex and age were adjusted for (for FM and BMI, Figures 1 and 2, respectively). The corresponding ANCOVA results for %BF were sex, P = 0.474; age group, P = 0.221; GC, P < 0.001; GC-by-sex interaction, P < 0.001; GC–by–age group interaction, P = 0.01. In a further ANCOVA including all children for whom maternal BMI data were available (n = 196), the relation between glucocorticoids and body composition was examined after GC was related to individual BSA. This was done because the adrenal gland’s size and cortisol secretion are closely correlated with BSA. The association between GC and body composition remained significant and interactions with age and sex were no longer significant (Table 2).
View this table:
TABLE 2 . Covariance analysis with age group and sex as grouping variables and GC [the sum of the 3 major urinary glucocorticoid metabolites; corrected for body surface area (BSA)] and mother’s BMI as covariates1
DISCUSSION
The average daily excretion rates of cortisol and its metabolites observed in the present investigation were in the same range as the physiologic steroid output values reported in other studies in children (25–28). However, these studies did not specifically examine the relation between glucocorticoid secretion and body composition. The mean values for BMI in the different sex and age groups in the present study corresponded closely with recently published values for a German reference population of children and adolescents (29). Also, the birth weights of the children in our study (n = 295 for whom was data available; boys: 3532 ± 500 g; girls: 3334 ± 457 g) were in the usual range of healthy subjects.
Our hormonal and anthropometric findings obtained in a large sample of healthy, primarily nonoverweight children confirm earlier reports (on small groups of obese children and control subjects), which indicated positive associations between endogenous GC excretion and fat mass (11, 12). In our girls, 10–32% of the variability of body fat variables could be explained by variation in adrenal glucocorticoid output (simple linear regression; Figures 1 and 2). In boys, the associations were less clear, significant only in late prepuberty (FM, BMI) and puberty (BMI). The reason for this sex difference is not known; however, after GC was corrected for individual BSA and was further adjusted for the effect of maternal BMI, the interaction between sex and GC was no longer significant (Table 2).
In contrast to GC, no associations were discernible for UFF and UFE, which, in principle, is in accordance with the findings of Knutsson et al (9). They found no relation between average 24-h plasma cortisol and the weight-for-height index in healthy children at different pubertal stages. Unfortunately, Knutsson et al did not measure more direct variables of body fatness. Furthermore, the concentrations of total plasma cortisol (as measured by Knutsson et al) do not necessarily reflect the free (tissue-available) fraction, because a high percentage of cortisol is bound to corticosteroid-binding globulin, and the concentration of circulating corticosteroid-binding globulin is partly genetically determined (30). On the other hand, it can be assumed that 24-h excretion rates of UFF and UFE primarily reflect an integrative measure of those glucocorticoids with biological activity (cortisol and cortisone) that are not bound to corticosteroid binding globulin. Together, Knutsson et al’s results and our cross-sectional study suggest that during normal growth immediate effects of cortisol (within the physiologic range of variation) appear not to play a primary role in body fatness.
However, the above-mentioned association between indexes of body fatness and 24-h urinary GC excretion (which reflects overall adrenal cortisol secretion) remained significant even after GC was corrected for BSA and the effect of maternal BMI on children’s BMI (14, 15) was adjusted for. BSA-normalized GC excretion proved to be a predictor of BMI during growth. Because no such association was discernible for UFF and UFE, one could argue that cortisol metabolism may be enhanced in children with higher body fat and higher overall cortisol secretion. As a result of an enhanced metabolic clearance of cortisol, there would be less negative feedback suppression of adrenocorticotropin-dependent corticosteroid secretion, leading to an overall increased GC output. Several authors have argued that such an increase in cortisol metabolism could be caused by a reduced hepatic 11ß-HSD1 activity, resulting in an impaired reactivation of cortisone to cortisol (31–33). However, although 11ß-HSD1 (as reflected by the ratio of THF + 5-THF/THE) significantly varied between the age groups, it did not correlate with any index of fatness, except in pubertal boys (negative association with %BF). This principal failure to detect a correlation between 11ß-HSD1 and body fat might be caused by a compensation of a down-regulated 11ß-HSD1 in the liver (eg, by an increased insulin-like growth factor 1 bioactivity; 34, 35) with an up-regulated 11ß-HSD1 in subcutaneous or visceral adipose tissue or both (36, 37). As recently discussed in detail by Rask et al (31), such a mechanism could explain a large number of conflicting reports on urinary cortisol-cortisone metabolite ratios in human obesity.
Also, no associations between 11ß-HSD2 (UFF/UFE) and indexes of body fat mass were discernible. Thus, it was not possible in the present study to identify a particular enzyme activity that could be responsible for the elevated cortisol metabolism in children with higher body fat. Whether the observed higher adrenocortical activity in fatter (predominantly nonobese) children may be secondarily caused by body fatness itself remains unclear. For the major secretion product of fat tissue, leptin, both inhibitory (38–40) and stimulatory (41) effects on adrenocortical activity have been described.
Taken together, our results confirm, in a large number of healthy children and adolescents, the existence of an association between endogenous glucocorticoid secretion and body composition. Even after urinary GC output was corrected for individual BSA and was further adjusted for the effect of maternal BMI, a significant portion of the variability in body fatness could be explained by variations in GC excretion. No associations with indexes of body fat were seen for 24-h excretion rates of UFF and UFE or for 11ß-HSD1 (except in pubertal boys) or 11ß-HSD2 activity. Thus, an immediate effect of physiologically varying cortisol or cortisone on body composition could not be confirmed. Our findings suggest that a higher adrenocortical activity is one endocrine-metabolic feature of healthy children with higher body fat. Whether this elevated adrenocortical activity is genetically determined or caused by fetal programming and whether it is causally involved in long-term fat accretion during growth should be examined in future longitudinal studies.
ACKNOWLEDGMENTS
TD assisted in the design of the study and took primary responsibility for data analysis and writing the paper. CM-G organized and supervised the steroid hormone analyses and provided scientific advice for the final results. TR designed the study and contributed substantially to the writing of the manuscript and the interpretation of the findings. None of the authors has any conflict of financial or personal interest in connection with this article.
REFERENCES