| HOME | HELP | FEEDBACK | SUBSCRIPTIONS | ARCHIVE | SEARCH | TABLE OF CONTENTS |
Departments of Pediatrics (J.Z.K.-V., A.R., T.M.), Biostatistics (W.Y.), and Movement Science (K.T.B.), University of Michigan, Ann Arbor, Michigan 48109
Address all correspondence and requests for reprints to: Josephine Z. Kasa-Vubu, M.D., M.S., Department of Pediatrics, University of Michigan Medical Center, Ann Arbor, Michigan 48019-0718.
| Abstract |
|---|
|
|
|---|
Methods: We measured body composition and maximal oxygen consumption (VO2max) in 37 postpubertal adolescent girls aged 1621 yr. GH was sampled every 10 min and leptin hourly for 24 h. We first analyzed 6-h time blocks by repeated measures for GH and leptin, with body mass index (BMI), percent body fat, and VO2max as covariates for the entire group and a lean subgroup. The deconvolution method was used to characterize GH pulsatility from individual time points.
Results: GH varied through the day (P < 0.0001), with the highest concentrations overnight. BMI, percent body fat, and VO2max were related to GH concentrations in the entire group, whereas leptin predicted GH in the entire group as well as the lean subgroup of girls. Higher leptin was related to lower GH concentrations (P = 0.011), regardless of time. A log leptin level increase by 1 unit decreased GH by 27%. Pulsatility characteristics showed a 1-yr increase of age at menarche increasing total GH input by 20% (P = 0.0035) independently from BMI.
Conclusion: In postpubertal adolescent girls, leptin is related to GH concentration across the lean to overweight BMI spectrum. GH pulsatile secretion was greater in girls with later age at menarche.
| Introduction |
|---|
|
|
|---|
The parallelism between GH and leptin, a marker of fat mass (5), has been described in various populations of obese and nonobese subjects (3, 4). Several recent studies also highlighted the link between GH and metabolic indices related to weight control such as visceral abdominal fat, insulin (6), leptin, and ghrelin (7) in adults and elderly subjects. Cardiovascular fitness is also linked to exercise and weight control.
There is a rapidly declining rate of physical activity in adolescent girls (8) as well as rising rates of obesity in women (9). In contrast to this, regular, intense exercise is associated with endocrine disruptions in adolescent girls. A state of negative energy balance as found with athletic training has been linked to a relative decrease in leptin in young women (10). Whereas athleticism in lean girls has been extensively studied (11, 12, 13), sedentary or moderately active adolescent girls have been the focus of considerably fewer studies.
Leptin is an adipocyte-derived hormone that is related to body mass index (BMI) but is also negatively impacted by fitness and can be used as an index of energy balance (10). We thus hypothesized that leptin is related to GH secretion in healthy postpubertal adolescent girls across the weight and activity spectra.
| Subjects and Methods |
|---|
|
|
|---|
The protocol was approved by the Institutional Review Board of the University of Michigan Health System. Thirty-seven healthy adolescents aged 1621 yr with a BMI between the 10th and 95th percentile for age (based on growth charts issued by the National Institute for Health Statistics) were studied a minimum of 2 yr after the age at menarche as part of a study on exercise and the menstrual cycle. Because the range of menarche in the United States is 1116 yr (14), onset of menarche at 16 yr was selected as the upper limit for inclusion in the study. Young women were recruited from local university campuses and high schools and were enrolled after signing an institutional review board-approved consent form. An additional signature was required from one parent or legal guardian for all participants under 18 yr of age. All participants were healthy and taking no medications. Individuals who had used any hormonal method of birth control or reported a significant weight loss (more than 10%) within the last year were excluded. A 3-d prospective food diary was collected and analyzed. All girls were asked to estimate, if any, how many hours of exercise they practiced every week (reported exercise) and describe the type of exercise. Self-described sedentary girls who reported no planned physical activity were not excluded. All participants attended classes on a college or high school campus, and all walked from class to class. In addition, we characterized fitness with the measure of maximal aerobic capacity (VO2max) as determined by a treadmill test following the Bruce protocol (15). Relative body fat was obtained by dual-energy x-ray absorptiometry using a total body scanner (model DPX-L; Lunar Radiation Corp., Madison, WI) (16). As a screen for potential causes of pituitary dysfunction, TSH and prolactin were assessed in all adolescents. Pregnancy was excluded by serum measurement of human chorionic gonadotropin.
All participants with an established menstrual cycle were instructed to contact the principal investigator at the start of the menstrual cycle and were admitted to the General Clinical Research Center of the University of Michigan within the first 10 d of the cycle. The follicular phase of the cycle was further confirmed by a serum progesterone measured on the night preceding the study and before proceeding with sample collection. Girls with irregular menses were admitted at random, with the provision that the prestudy progesterone had to be at a follicular phase level, to proceed with data collection. This strategy allowed to consistently target the follicular phase of the cycle (average serum progesterone of 0.4 ± SD 0.2 ng/ml). The admission to the General Clinical Research Center was also timed on the night before the study to allow for acclimatization. By 0700 h, an iv catheter was inserted into the forearm, and blood was sampled 24 h for GH every 10 min as well as leptin every hour.
Assays
Plasma samples were stored at 70 C until assayed. Plasma GH concentrations were measured using a chemiluminescent enzyme immunometric assay for use with the Immulite automated analyzer (Diagnostic Products Corp., Los Angeles, CA). Assay sensitivity was 0.01 ng/ml. The intra- and interassay coefficients of variation (CVs) were 2.0 and 4.5%, respectively. Leptin was measured using reagents from Linco Research (St. Charles, MO) with intra- and interassay CVs of 4.5 and 6.6%. To further characterize the GH axis (17), we measured IGF-I using a chemiluminescent enzyme immunometric assay for use with the Immulite automated analyzer (Diagnostic Products). The assay sensitivity was 20 ng/ml; intra- and interassay CVs were 2.5 and 10.7%, respectively.
Analysis
The analysis applied two different approaches to first describe more general patterns throughout the day, focused on large time blocks, and then further detailing pulsatile characteristics of GH derived from frequent individual samples. We took into account fitness as measured by maximal oxygen consumption, fatness as measured by total body percent body fat, and age at menarche.
Time-block means analysis. As a first step, we delineated 6-h time blocks to see whether there were relationships that could be related to integrated windows of time reflecting different times of the day (overnight, early morning, midday, and evening). Because of the nighttime prominence of GH in youth, 2300 h was arbitrarily chosen as the beginning of the first time block or time 1. The 145 time points of GH and 25 hourly leptin measurements were analyzed as four 6-h blocks defined as follows: time 1 = 23000450 h, time 2 = 05001050 h; time 3 = 11001650 h; time 4 = 17002250 h. The mean concentration of both hormones for each time interval was then calculated and referred to as block mean. Repeated-measures ANOVA was used to determine the effects of time as well as the effect of body composition, VO2 max, and age at menarche on GH and leptin. A Tukey adjustment was used for post hoc analysis of different time periods. Each time block mean was log (base e) transformed before analysis.
Pulsatility analysis. As a second step, we analyzed the pulsatile characteristics of GH over 24 h. Deconvolution (18) was fitted to each series to detect and characterize the pulsatility in GH. Total input was defined as the sum of area under all the recognized GH pulses. Half-life, mean area under the curve, total input, mean height, and number of pulses were the summary measures of interest. Linear regression was used to study the relationship among VO2max, BMI, and percent body fat on GH. All analyses were adjusted for age at menarche. Half-life, mean area under the curve, and total input were log (base e) transformed before analysis, and the number of pulses was square root transformed before analysis.
| Results |
|---|
|
|
|---|
|
We examined relationships between leptin as an index of energy balance and weight, adiposity, and fitness. Regardless of time of collection, a 1% increase in body fat related to a 7% increase of leptin (P < 0.0001), whereas a 1-unit increase in BMI increased leptin by 15% (P < 0.0001), and a 1-unit increase in VO2max decreased leptin by 6% (P < 0.0001). When both percent body fat and VO2max were studied together, percent body fat was still significantly related to leptin (P < 0.0001), but the significance of VO2max disappeared (P = 0.69). However, BMI and VO2max both retained significance within the same model for leptin (P = 0.0032 and 0.0154, respectively). Thus, the relationship between VO2max and leptin is not mediated by BMI changes but appears dependent on differences in percent body fat.
GH Time block analysis
There was marked diurnal variation for GH during the 6-h time block, with the highest concentrations reached overnight between 2300 and 450 h as outlined in Table 2
.
|
Univariate relationships between GH and BMI and percent body fat and VO2max for the entire group are shown in Fig. 1
. VO2max, BMI, and percent body fat were all highly related. The Pearson coefficients of correlation were 0.87 for BMI and percent body fat, 0.58 for BMI and VO2max, and 0.75 for percent body fat and VO2max. As a result, none of the three variables emerged as a predominant predictor of GH when analyzed in a multiple regression model.
|
Leptin showed a marked diurnal variation with a peak between 2300 and 0450 h and a nadir in the afternoon followed by an evening rise (Table 3
). Overall, higher leptin concentrations were related to lower GH concentrations (P = 0.011) and a log leptin level increase by 1 unit related to a mean GH decrease of 27%. This relationship was maintained regardless of the time block of observation. This relationship was also maintained when the analysis was restricted to the leanest girls with a BMI between 18 and 25 kg/m2 (P value for leptin = 0.015). Within the lean subgroup, a log leptin level increase by 1 unit related to a mean GH decrease of 26%. The correlation of leptin with weight and fitness indices was strong, and there was no independent effect of either variable on mean GH concentration.
|
We studied five pulsatility parameters of GH: half-life, mean area under the curve, total input, input, and mean peak amplitude. Representative profiles are shown in Fig. 2
. Half-life, number of pulses, and mean amplitude of GH were not significantly related to VO2max, BMI, or percent body fat (results not shown). The total input or pulse-related GH was linked to age at menarche (P = 0.0035). For every additional year at menarche GH total input increased by 20%. The total input of GH was not related to VO2max (P = 0.16) or BMI (P = 0.11); however, there was a negative trend with percent body fat that did not reach statistical significance (P = 0.066).
|
| Discussion |
|---|
|
|
|---|
Although leptin was closely related to the other surrogate measures of energy balance, namely BMI, VO2max, and percent body fat, it was associated to GH mean concentration in the subgroup of lean girls with a BMI between 18 and 25 kg/m2. Our study further illustrates the limitations of BMI as a weight measure and/or a surrogate measure of adiposity in young subjects who can represent a wide range of physical fitness and body composition. The importance of fitness is likely to be more evident in youth because baseline activity tends to be greater than in older adults. Analyzed together, both BMI and VO2max remained significantly related to leptin. We propose that in obesity-related studies in adolescent and young women, the assessment of BMI should be complemented by a measure of adiposity as well as a measure of fitness or endurance to exercise.
Although all girls were studied a minimum of 2 yr after menarche, there was higher GH secretion with menarche at an older age. Because GH decreases after puberty (25) and declines further with aging (26), one possible explanation is that the girls with a later age at menarche were still immature with stronger GH pulses, as can be seen in younger individuals. Our study highlights that the postpubertal maturation of neuroendocrine function continues for at least longer than 23 yr. Because exercise habits are acquired over a long period of time (27), one can speculate that girls with a later age at menarche were likely to be more active as they entered puberty, thus delaying their menarche. This would in part explain why higher VO2max was related to significantly higher GH secretion. Because adolescents with delayed puberty may also have a more prolonged course until they achieve maturation, it is also possible that a slower tempo of puberty would influence the results in girls with a later age at menarche.
In recent years, there have been reports suggesting that GH is involved in pubertal maturation (28), facilitates folliculogenesis (29), and enhances the effectiveness of assisted reproductive techniques in some women (30). Attempts to discriminate between women in which GH was effective at facilitating ovulation from those who were not GH responsive have relied on the response to GH-provocative tests as predictors (31, 32). Salat-Baroux et al. (33) reported that basal GH levels were higher in women who were poor responders to assisted ovulation techniques than those who had a good response. Due to our participants young age and the cross-sectional nature of this study, fertility potential could not be assessed here. It is possible, however, that the concurrent finding of higher GH secretion and decreased fertility merely represents the impact of acquired fitness during puberty by delaying puberty and fertility and enhancing GH secretion.
| Acknowledgments |
|---|
| Footnotes |
|---|
The authors have nothing to declare.
First Published Online July 25, 2006
Abbreviations: BMI, Body mass index; CV, coefficient of variation; VO2max, maximal oxygen consumption.
Received December 30, 2005.
Accepted July 18, 2006.
| References |
|---|
|
|
|---|
This article has been cited by other articles:
![]() |
J. Z. Kasa-Vubu, A. Rosenthal, E. G. Murdock, and K. B. Welch Impact of Fatness, Fitness, and Ethnicity on the Relationship of Nocturnal Ghrelin to 24-Hour Luteinizing Hormone Concentrations in Adolescent Girls J. Clin. Endocrinol. Metab., August 1, 2007; 92(8): 3246 - 3252. [Abstract] [Full Text] [PDF] |
||||
| |||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||
| HOME | HELP | FEEDBACK | SUBSCRIPTIONS | ARCHIVE | SEARCH | TABLE OF CONTENTS |
| Endocrinology | Endocrine Reviews | J. Clin. End. & Metab. |
| Molecular Endocrinology | Recent Prog. Horm. Res. | All Endocrine Journals |