Add Journal to My Library
Journal of Clinical Endocrinology and Metabolism
, Volume 103 (8) – Aug 1, 2018

10 pages

/lp/ou_press/ghrelin-and-peptide-yy-change-during-puberty-relationships-with-IvMbMR3w4k

- Publisher
- Endocrine Society
- Copyright
- Copyright © 2018 Endocrine Society
- ISSN
- 0021-972X
- eISSN
- 1945-7197
- D.O.I.
- 10.1210/jc.2017-01825
- Publisher site
- See Article on Publisher Site

Abstract Context Pubertal adolescents show strong appetites. How this is mediated is unclear, but ghrelin and peptide YY (PYY) play potentially important roles. Objective To measure ghrelin and PYY change in relation to pubertal growth. Design Three-year prospective cohort study. Setting Australian regional community. Participants Eighty healthy adolescents (26 girls; 54 boys) recruited at 10 to 13 years. Main Outcome Measures Fasting circulating total ghrelin, total PYY, IGF-1, insulin, leptin (via radioimmunoassay), estradiol and testosterone (via mass spectrometry), anthropometry, and body composition (via bioelectrical impedance). Results Adolescents exhibited normal developmental change. Mixed models revealed positive associations for ghrelin to age2 (both sexes: P < 0.05), indicating a U-shaped trend over time. Ghrelin was also inversely associated with IGF-1 (both sexes: P < 0.05), leptin in girls (P < 0.01), and insulin in boys (P < 0.05) and negatively correlated with annual height and weight velocity (both sexes: P ≤ 0.01). PYY showed no age-related change in either sex. Neither ghrelin nor PYY were associated with Tanner stage. Weight subgroup analyses showed significant ghrelin associations with age2 in healthy-weight but not overweight and obese adolescents (7 girls; 18 boys). Conclusions Adolescents showed a U-shaped change in ghrelin corresponding to physical and biochemical markers of growth, and no change in PYY. The overweight and obesity subgroup exhibited an apparent loss of the U-shaped ghrelin trend, but this finding may be attributed to greater maturity and its clinical significance is unclear. Further research on weight-related ghrelin and PYY trends at puberty is needed to understand how these peptides influence growth and long-term metabolic risk. During puberty, adolescents exhibit a rise in estradiol or testosterone (the key puberty hormones in girls and boys, respectively), rapid linear growth, and sex-specific patterns of lean and fat mass deposition (1–3). To fuel such rapid physical change, energy requirements must increase at puberty. Evidence from 20 years of cross-sectional research has shown that energy expenditure (and intake) in pubertal adolescents are higher than their prepubertal counterparts (4). Yet, data from longitudinal studies show apparently paradoxical outcomes (5, 6). In the Early Bird Diabetes cohort study, a reduction in absolute energy expenditure was observed across the age span of 11 to 15 years in 320 adolescents (5). This reduction suggests an increase in energy efficiency, which was speculated as an adaptive response to the nutritional pressures of growth (5) and affirms the importance of longitudinal studies for understanding energy balance at puberty. Energy intake is integral to energy balance and in adults is regulated via neuronal circuits interacting with appetite hormones, key among these being ghrelin and peptide YY (PYY) (7). Ghrelin is a hormone well recognized for its acute orexigenic properties (8). Yet, ghrelin has wider physiological roles that include promoting GH release and maintaining glucose homeostasis (8). Increasing evidence also indicates an involvement of ghrelin in signaling negative energy balance and in modulating reproductive function (9). Conversely, PYY is best known for its anorexigenic effects. The peptide is responsible for inducing satiety and cessation of eating after food intake (7) and plays a potential role in the long-term regulation of nutrient status, although current evidence is inconsistent (10, 11). Studies have reported lower circulating PYY with increasing levels of GH (12, 13). Thus, alterations in circulating ghrelin and PYY influence the physiological drive to eat, weight gain, and reproductive function. These factors make study of these hormones highly relevant at puberty, when rapid growth and maturation must require careful coordination of energy balance and appetite regulatory signals (14). Dysregulation of such signals may characterize the recognized risk of development of overweight and obesity (Ow/Ob) in adolescents (14, 15). There is a paucity of longitudinal data on pubertal changes in ghrelin and PYY. One cohort study in female rhythmic gymnasts reported a decline in fasting ghrelin levels from prepuberty to early puberty, a trend apparently contradictory to its orexigenic effects, but also consistent with its other physiological roles, and with prior cross-sectional data (16–19). Current literature on pubertal PYY change is limited to two cross-sectional studies with conflicting findings (13, 20). A research gap thus remains in understanding how ghrelin and PYY change across this life stage, and the relationship these peptides have with other aspects of pubertal growth and development. The primary aim of this study was to measure change in ghrelin and PYY in adolescents, and understand how these peptides relate to the physical and biochemical markers of pubertal growth and development. A secondary aim was to compare ghrelin and PYY between healthy-weight adolescents and those with Ow/Ob. We hypothesized, based on cross-sectional data, that circulating ghrelin and PYY levels would decrease with the pubertal growth spurt. Adolescents with Ow/Ob were expected to show lower ghrelin and PYY levels, reflecting a state of energy surplus but a reduced capacity for satiety signaling. Subjects and Methods Study setting and participants Participants came from the longitudinal Adolescent Rural Cohort study of Hormones, health, Education, environments and Relationships (ARCHER). ARCHER is a prospective cohort study of 342 healthy adolescents, recruited between mid-2011 and mid-2013 at the age of 10 to 13 years (21), that aims to examine how change in puberty hormones affect health and wellbeing in young people. This ongoing study collects data from families living in two Australian regional townships, and with up to 5 years of data available. Data collection includes annual anthropometry, blood collection, and a health questionnaire (22). A total of 80 ARCHER participants (26 girls and 54 boys) were selected for ghrelin and PYY analyses. We selected those participants who provided the greatest number of serial plasma samples (out of a possible five) over the follow-up period. Where there was an equal number of samples available from multiple participants, priority was given to those who provided plasma at the earlier follow-ups (i.e., at the less advanced stages of puberty) (Supplemental Fig. 1). Ethics Ethical approval was granted by the Human Research Ethics Committee of the University of Sydney (HREC 2010/13094 and 2015/199). All research was undertaken in accordance with the Principles of the Declaration of Helsinki. Written informed consent was obtained from adolescents and their parent/guardian prior to study commencement. Adolescents provided verbal assent prior to each study visit and blood collection. Anthropometric and body composition measurement Height (to the nearest 0.1 cm) was measured with a stadiometer (Wedderburn Pty Ltd, Ingleburn, Australia). Weight (to the nearest 0.1 kg) and body composition were measured in light clothing and no shoes using a foot-to-foot bioelectrical impedance scale (Tanita Corporation, Tokyo, Japan) (23). Height and weight velocity were calculated as the difference in height/weight divided by the difference in age between two consecutive visits, and assigning this value to the age at the latter visit. Body mass index (BMI) z scores and Ow/Ob status were derived using the extended International Obesity Task Force cutoffs (24). Waist circumference was measured at the narrowest point, to the nearest 0.1 cm. The waist-to-height ratio was calculated as an indicator of central adiposity (25). Pubertal assessment Self-reported Tanner stage was assessed as a part of the annual health questionnaire. Adolescents were asked to rate their pubertal status against standardized line drawings (26, 27). Girls were asked if/when they had reached menarche. Biochemistry A fasting morning venous blood sample was collected annually (cycle day 7 to 10 in postmenarcheal girls). Serum was collected at all annual time points. Plasma was collected only from the first year of follow-up [i.e., no baseline plasma sample, using vacutainers with EDTA and aprotinin (BD, North Ryde, Australia)]. Samples were centrifuged at 4°C (1600g × 15 minutes) immediately after collection, aliquoted, and stored at –80°C. Serum was used for assay of estradiol, testosterone, glucose, insulin, and leptin. Plasma was used for assay of total ghrelin and total PYY. Estradiol and testosterone were analyzed via liquid chromatography tandem mass spectrometry using the API 5000 System (AB SCIEX, Mount Waverley, Australia); assay performance characteristics are published (28). Glucose was measured using a glucose hexokinase enzymatic assay on the AU480 chemistry analyzer (Beckman Coulter, Lane Cove, Australia). Commercial radioimmunoassay kits were used to measure IGF-1 (Mediagnost, Reutlingen, Germany), insulin, leptin, total ghrelin, and total PYY (Merck Millipore, Billerica, MA). No apparent cross-reactivity is reported for the IGF-1 kit, and 100% specificities are reported in the respective kits for human insulin, human leptin, desacyl and acyl-ghrelin (i.e., total ghrelin), and PYY 1-36 and 3-36 (i.e., total PYY). Mean interassay coefficients of variation were 5.1% for IGF-1, 3.9% for insulin, 4.5% for leptin, 16.3% for total ghrelin, and 7.0% for total PYY. Mean intra-assay coefficients of variation were 4.8% for IGF-1, 3.6% for insulin, 3.1% for leptin, 3.9% for total ghrelin, and 2.9% for total PYY. The updated homeostasis model assessment-insulin resistance (HOMA2-IR) was calculated as a measure of insulin resistance (29). Conversion from Système International to conventional units is as follows: estradiol pmol/L ÷ 3.671 = pg/mL; testosterone nmol/L ÷ 3.467 = ng/mL; IGF-1 nmol/L ÷ 0.131 = ng/mL; glucose mmol/L ÷ 0.055 = mg/dL; insulin pmol/L ÷ 6.945 = µIU/mL; total ghrelin pmol/L ÷ 0.296 = pg/mL; and PYY pmol/L ÷ 0.25 = pg/mL. Statistical analysis Sample size calculations using cross-sectional data indicated the study had adequate power to detect changes in ghrelin and PYY over time and across Ow/Ob subgroups (Supplemental Materials and Methods). To identify potential selection bias in the current subcohort of the ARCHER study, baseline characteristics of included vs excluded participants were compared using sex-specific unpaired t tests and χ2 tests. Data are reported as mean ± SD or median (interquartile range). All descriptive analyses were performed using IBM SPSS Statistics version 24 (IBM Corporation, Armonk, NY). To describe how markers of growth and development changed over time, anthropometric and biochemical (estradiol, testosterone, insulin, HOMA2-IR, IGF-1, and leptin) data were regressed against age (in fractional years) and Tanner stage using sex-specific mixed models with random intercepts. Functional forms of age and Tanner stage were tested including linear, quadratic, cubic, log, and square root to assess the most appropriate relationship as determined by significance of the terms (P < 0.05). Sexual dimorphism was assessed by running the above models on the combined female and male data set, with sex as an additional predictor. Anthropometric and biochemical changes were plotted against age (rounded to the nearest year) and Tanner stage and presented graphically as mean ± SE. The relationship of ghrelin and PYY to age (fractional years), anthropometry, and growth-related biomarkers was analyzed using a supervised forward stepwise procedure in a mixed-model framework with random intercepts. For the first step, ghrelin and PYY were included as separate outcomes and regressed against age, which was the main time-related exposure variable (model 1). Again, various functional forms of age were tested to identify the most appropriate relationship as determined by significance of the terms (P < 0.05). As ghrelin showed a quadratic relationship over time, the timing of the ghrelin nadir was assessed by comparing model-derived estimates of age at ghrelin nadir and age at peak height velocity via unpaired t tests. Spearman’s correlation was also used to examine the association of ghrelin to height and weight velocity. The second step in mixed modeling involved testing a list of potential predictors including Tanner stage (in its various functional forms), height, weight, BMI z score, waist circumference, fat mass, fat-free mass, insulin, leptin, estradiol in girls, testosterone in boys, PYY (in models with ghrelin as an outcome), and ghrelin (in models with PYY as an outcome). Each of these potential predictors were analyzed separately (but included age terms and IGF-1) against ghrelin and PYY and were retained in a final model (model 2) if it was the most significant predictor, as determined by the magnitude of the P value less than 0.05, until there were no more significant variables or when multicollinearity was present (measured by a variance inflation factor >2). Models 1 and 2 were applied separately to healthy-weight and Ow/Ob subgroups to examine the weight-related differences in predictors of ghrelin and PYY. Outcomes from mixed modeling are reported as estimates, P values, and 95% CIs. All statistical modeling was performed using PROC MIXED in SAS version 9.4 (SAS Institute, Cary, NC). Results Baseline age was 11.5 ± 0.9 years, and baseline height and weight were 149.8 ± 9.0 cm and 39.9 (25.5 to 54.3) kg (Table 1). Prevalence of Ow/Ob was 31.3% (n = 25; 7 girls and 18 boys). Compared with the remaining 262 ARCHER adolescents, girls included in the current study were younger by 7 months (P = 0.007), and boys had marginally lower fat-free mass (by 0.1 kg; P = 0.047) and serum estradiol (by 0.5 pmol/L; P = 0.001). All other baseline variables, including anthropometry, body composition, Ow/Ob prevalence, and puberty hormones, were similar between the included and excluded participants. Table 1. Baseline Characteristics of Adolescents Included in the Current Study Baseline Characteristic All Participants (N = 80) Girls (n = 26) Boys (n = 54) Age, y 11.5 ± 0.9 11.2 ± 0.8 11.6 ± 0.9 Height, cm 149.8 ± 9.0 149.2 ± 7.5 150.1 ± 9.7 Weight, kg 39.9 (25.5–54.3) 41.1 (28.5–53.7) 38.7 (23.3–54.1) BMI z score 0.61 ± 1.03 0.77 ± 1.06 0.53 ± 1.02 Overweight, % 20.0 11.5 24.1 Obese, % 11.3 15.4 9.3 Waist circumference, cm 63.6 (52.0–75.2) 63.8 (55.4–72.2) 63.4 (51.4–75.4) Waist-to-height ratio 0.44 ± 0.06 0.45 ± 0.07 0.44 ± 0.05 Body fat, % 20.8 ± 10.3 24.3 ± 9.4 19.1 ± 10.4 Fat mass, kg 7.2 (0.7–13.7) 9.0 (0.0–18.0) 9.4 (4.0–14.8) Fat-free mass, kg 32.5 (25.4–39.6) 32.5 (26.3–38.7) 32.7 (26.0–39.4) Testosterone, nmol/L 0.45 (0.0–2.43) 0.39 (0.0–1.08) 0.80 (0.0–4.30) Estradiol, pmol/L 38.9 (0.0–115.2) 93.1 (0.0–286.8) 30.1 (0.0–76.3) Baseline Characteristic All Participants (N = 80) Girls (n = 26) Boys (n = 54) Age, y 11.5 ± 0.9 11.2 ± 0.8 11.6 ± 0.9 Height, cm 149.8 ± 9.0 149.2 ± 7.5 150.1 ± 9.7 Weight, kg 39.9 (25.5–54.3) 41.1 (28.5–53.7) 38.7 (23.3–54.1) BMI z score 0.61 ± 1.03 0.77 ± 1.06 0.53 ± 1.02 Overweight, % 20.0 11.5 24.1 Obese, % 11.3 15.4 9.3 Waist circumference, cm 63.6 (52.0–75.2) 63.8 (55.4–72.2) 63.4 (51.4–75.4) Waist-to-height ratio 0.44 ± 0.06 0.45 ± 0.07 0.44 ± 0.05 Body fat, % 20.8 ± 10.3 24.3 ± 9.4 19.1 ± 10.4 Fat mass, kg 7.2 (0.7–13.7) 9.0 (0.0–18.0) 9.4 (4.0–14.8) Fat-free mass, kg 32.5 (25.4–39.6) 32.5 (26.3–38.7) 32.7 (26.0–39.4) Testosterone, nmol/L 0.45 (0.0–2.43) 0.39 (0.0–1.08) 0.80 (0.0–4.30) Estradiol, pmol/L 38.9 (0.0–115.2) 93.1 (0.0–286.8) 30.1 (0.0–76.3) Data are presented as mean ± SD, median (interquartile range), or a percentage. View Large Table 1. Baseline Characteristics of Adolescents Included in the Current Study Baseline Characteristic All Participants (N = 80) Girls (n = 26) Boys (n = 54) Age, y 11.5 ± 0.9 11.2 ± 0.8 11.6 ± 0.9 Height, cm 149.8 ± 9.0 149.2 ± 7.5 150.1 ± 9.7 Weight, kg 39.9 (25.5–54.3) 41.1 (28.5–53.7) 38.7 (23.3–54.1) BMI z score 0.61 ± 1.03 0.77 ± 1.06 0.53 ± 1.02 Overweight, % 20.0 11.5 24.1 Obese, % 11.3 15.4 9.3 Waist circumference, cm 63.6 (52.0–75.2) 63.8 (55.4–72.2) 63.4 (51.4–75.4) Waist-to-height ratio 0.44 ± 0.06 0.45 ± 0.07 0.44 ± 0.05 Body fat, % 20.8 ± 10.3 24.3 ± 9.4 19.1 ± 10.4 Fat mass, kg 7.2 (0.7–13.7) 9.0 (0.0–18.0) 9.4 (4.0–14.8) Fat-free mass, kg 32.5 (25.4–39.6) 32.5 (26.3–38.7) 32.7 (26.0–39.4) Testosterone, nmol/L 0.45 (0.0–2.43) 0.39 (0.0–1.08) 0.80 (0.0–4.30) Estradiol, pmol/L 38.9 (0.0–115.2) 93.1 (0.0–286.8) 30.1 (0.0–76.3) Baseline Characteristic All Participants (N = 80) Girls (n = 26) Boys (n = 54) Age, y 11.5 ± 0.9 11.2 ± 0.8 11.6 ± 0.9 Height, cm 149.8 ± 9.0 149.2 ± 7.5 150.1 ± 9.7 Weight, kg 39.9 (25.5–54.3) 41.1 (28.5–53.7) 38.7 (23.3–54.1) BMI z score 0.61 ± 1.03 0.77 ± 1.06 0.53 ± 1.02 Overweight, % 20.0 11.5 24.1 Obese, % 11.3 15.4 9.3 Waist circumference, cm 63.6 (52.0–75.2) 63.8 (55.4–72.2) 63.4 (51.4–75.4) Waist-to-height ratio 0.44 ± 0.06 0.45 ± 0.07 0.44 ± 0.05 Body fat, % 20.8 ± 10.3 24.3 ± 9.4 19.1 ± 10.4 Fat mass, kg 7.2 (0.7–13.7) 9.0 (0.0–18.0) 9.4 (4.0–14.8) Fat-free mass, kg 32.5 (25.4–39.6) 32.5 (26.3–38.7) 32.7 (26.0–39.4) Testosterone, nmol/L 0.45 (0.0–2.43) 0.39 (0.0–1.08) 0.80 (0.0–4.30) Estradiol, pmol/L 38.9 (0.0–115.2) 93.1 (0.0–286.8) 30.1 (0.0–76.3) Data are presented as mean ± SD, median (interquartile range), or a percentage. View Large Growth over time Height, weight, and body composition changes over time were indicative of normal pubertal growth in our participants (Supplemental Figs. 2 and 3). Sexual dimorphism was observed for body composition trajectories, with girls and boys exhibiting significantly greater increases in fat and fat-free mass (P < 0.05), respectively, across both age and Tanner stage (Supplemental Figs. 2 and 3). Longitudinal biochemical trends Circulating levels of total ghrelin and total PYY were plotted against age and Tanner stage (Fig. 1). These plots showed a quadratic U-shaped trend over time for ghrelin, which was confirmed statistically by a significant and positive age2 effect (girls: P = 0.021; boys: P < 0.001) in the initial mixed models (Table 2, model 1). Functional forms of Tanner stage, including the linear (girls: P = 0.61; boys: P = 0.68) and quadratic (girls: P = 0.80; boys: P = 0.82) terms, did not significantly predict ghrelin over and above the strong age-related associations (Fig. 1). Model-derived estimates of mean age at ghrelin nadir and mean age at peak height velocity were, respectively, 13.3 and 12.7 years in girls and 14.4 and 13.9 years in boys. No significant difference was detected in the relative timing of these two events (girls: P = 0.21; boys: P = 0.48). When total ghrelin was correlated against annual height and weight velocity (Fig. 2), significant negative associations were observed, suggesting that the time of peak growth velocity was paralleled by a trough in circulating ghrelin levels. Figure 1. View largeDownload slide Sex-specific plots for (a) total ghrelin with age, (b) total ghrelin with Tanner stage, (c) total PYY with age, and (d) total PYY with Tanner stage, presented as mean ± SE. Mixed-model estimates and P values are provided for the quadratic and linear predictive terms, respectively, for the ghrelin and PYY plots. Observations by age in girls: 11 years: n = 3; 12 years: n = 10; 13 years: n = 17; 14 years: n = 21; 15 years: n = 12; and 16 years: n = 5. Observations by age in boys: 11 years: n = 2; 12 years: n = 13; 13 years: n = 31; 14 years: n = 39; 15 years: n = 34; and 16 years: n = 21. Observations by Tanner stage in girls: 1: n = 2; 2: n = 4; 3: n = 19; 4: n = 20; and 5: n = 12. Observations by Tanner stage (TS) in boys: 1: n = 3; 2: n = 21; 3: n = 22; 4: n = 29; and 5: n = 44. Figure 1. View largeDownload slide Sex-specific plots for (a) total ghrelin with age, (b) total ghrelin with Tanner stage, (c) total PYY with age, and (d) total PYY with Tanner stage, presented as mean ± SE. Mixed-model estimates and P values are provided for the quadratic and linear predictive terms, respectively, for the ghrelin and PYY plots. Observations by age in girls: 11 years: n = 3; 12 years: n = 10; 13 years: n = 17; 14 years: n = 21; 15 years: n = 12; and 16 years: n = 5. Observations by age in boys: 11 years: n = 2; 12 years: n = 13; 13 years: n = 31; 14 years: n = 39; 15 years: n = 34; and 16 years: n = 21. Observations by Tanner stage in girls: 1: n = 2; 2: n = 4; 3: n = 19; 4: n = 20; and 5: n = 12. Observations by Tanner stage (TS) in boys: 1: n = 3; 2: n = 21; 3: n = 22; 4: n = 29; and 5: n = 44. Table 2. Final Mixed-Effects Models Assessing Ghrelin and PYY Change in Relation to Physical and Biochemical Markers of Growth and Development, Analyzed by Sex Total Ghrelin (pmol/L) Total PYY (pmol/L) Girls (n = 26) Boys (n = 54) Girls (n = 26) Boys (n = 54) Estimate (95% CI) P Value Estimate (95% CI) P Value Estimate (95% CI) P Value Estimate (95% CI) P Value Model 1a Age, y –197.2 (–370.7, –23.8) 0.027 –228.9 (–256.8, –101.0) <0.001 1.86 (–0.54, 4.25) 0.127 −0.08 (–1.12, 1.29) 0.890 Age2 7.40 (1.16, 13.6) 0.021 7.94 (3.41, 12.5) <0.001 — — — — Model 2b Age, y −11.4 (–192.0, 169.3) 0.898 –150.9 (–271.8, –30.1) 0.015 2.07 (–0.12, 4.27) 0.064 0.31 (–1.28, 1.90) 0.698 Age2 0.49 (–6.20, 7.19) 0.881 5.57 (1.33, 9.81) 0.011 — — — — Waist, cm — — –2.30 (–4.25, –0.34) 0.022 –0.35 (–0.65, –0.05) 0.024 — — IGF-1, nmol/L –1.60 (–2.89, –0.30) 0.017 –0.92 (–1.50, –0.34) 0.002 0.17 (0.05, 0.30) 0.006 — — Insulin, pmol/L — — –0.21 (–0.38, –0.03) 0.022 0.05 (0.01, 0.09) 0.010 0.02 (–0.01, 0.04) 0.170 Leptin, nmol/L –33.5 (–55.9, –11.2) 0.004 — — — — 4.04 (0.07, 8.01) 0.046 Total Ghrelin (pmol/L) Total PYY (pmol/L) Girls (n = 26) Boys (n = 54) Girls (n = 26) Boys (n = 54) Estimate (95% CI) P Value Estimate (95% CI) P Value Estimate (95% CI) P Value Estimate (95% CI) P Value Model 1a Age, y –197.2 (–370.7, –23.8) 0.027 –228.9 (–256.8, –101.0) <0.001 1.86 (–0.54, 4.25) 0.127 −0.08 (–1.12, 1.29) 0.890 Age2 7.40 (1.16, 13.6) 0.021 7.94 (3.41, 12.5) <0.001 — — — — Model 2b Age, y −11.4 (–192.0, 169.3) 0.898 –150.9 (–271.8, –30.1) 0.015 2.07 (–0.12, 4.27) 0.064 0.31 (–1.28, 1.90) 0.698 Age2 0.49 (–6.20, 7.19) 0.881 5.57 (1.33, 9.81) 0.011 — — — — Waist, cm — — –2.30 (–4.25, –0.34) 0.022 –0.35 (–0.65, –0.05) 0.024 — — IGF-1, nmol/L –1.60 (–2.89, –0.30) 0.017 –0.92 (–1.50, –0.34) 0.002 0.17 (0.05, 0.30) 0.006 — — Insulin, pmol/L — — –0.21 (–0.38, –0.03) 0.022 0.05 (0.01, 0.09) 0.010 0.02 (–0.01, 0.04) 0.170 Leptin, nmol/L –33.5 (–55.9, –11.2) 0.004 — — — — 4.04 (0.07, 8.01) 0.046 a Model 1 assessed ghrelin and PYY change across time and includes age and age2 as fixed effects and participant ID and intercept as random effects. b Model 2 extended model 1 by testing additional predictors of ghrelin and PYY change using a forward stepwise approach. Additional predictors tested include Tanner stage, height, weight, BMI z score, waist circumference, fat mass, fat-free mass, IGF-1, insulin, leptin, testosterone in boys, and estradiol in girls. Results show only the predictors that significantly contributed to ghrelin or PYY change and were included in the final model. View Large Table 2. Final Mixed-Effects Models Assessing Ghrelin and PYY Change in Relation to Physical and Biochemical Markers of Growth and Development, Analyzed by Sex Total Ghrelin (pmol/L) Total PYY (pmol/L) Girls (n = 26) Boys (n = 54) Girls (n = 26) Boys (n = 54) Estimate (95% CI) P Value Estimate (95% CI) P Value Estimate (95% CI) P Value Estimate (95% CI) P Value Model 1a Age, y –197.2 (–370.7, –23.8) 0.027 –228.9 (–256.8, –101.0) <0.001 1.86 (–0.54, 4.25) 0.127 −0.08 (–1.12, 1.29) 0.890 Age2 7.40 (1.16, 13.6) 0.021 7.94 (3.41, 12.5) <0.001 — — — — Model 2b Age, y −11.4 (–192.0, 169.3) 0.898 –150.9 (–271.8, –30.1) 0.015 2.07 (–0.12, 4.27) 0.064 0.31 (–1.28, 1.90) 0.698 Age2 0.49 (–6.20, 7.19) 0.881 5.57 (1.33, 9.81) 0.011 — — — — Waist, cm — — –2.30 (–4.25, –0.34) 0.022 –0.35 (–0.65, –0.05) 0.024 — — IGF-1, nmol/L –1.60 (–2.89, –0.30) 0.017 –0.92 (–1.50, –0.34) 0.002 0.17 (0.05, 0.30) 0.006 — — Insulin, pmol/L — — –0.21 (–0.38, –0.03) 0.022 0.05 (0.01, 0.09) 0.010 0.02 (–0.01, 0.04) 0.170 Leptin, nmol/L –33.5 (–55.9, –11.2) 0.004 — — — — 4.04 (0.07, 8.01) 0.046 Total Ghrelin (pmol/L) Total PYY (pmol/L) Girls (n = 26) Boys (n = 54) Girls (n = 26) Boys (n = 54) Estimate (95% CI) P Value Estimate (95% CI) P Value Estimate (95% CI) P Value Estimate (95% CI) P Value Model 1a Age, y –197.2 (–370.7, –23.8) 0.027 –228.9 (–256.8, –101.0) <0.001 1.86 (–0.54, 4.25) 0.127 −0.08 (–1.12, 1.29) 0.890 Age2 7.40 (1.16, 13.6) 0.021 7.94 (3.41, 12.5) <0.001 — — — — Model 2b Age, y −11.4 (–192.0, 169.3) 0.898 –150.9 (–271.8, –30.1) 0.015 2.07 (–0.12, 4.27) 0.064 0.31 (–1.28, 1.90) 0.698 Age2 0.49 (–6.20, 7.19) 0.881 5.57 (1.33, 9.81) 0.011 — — — — Waist, cm — — –2.30 (–4.25, –0.34) 0.022 –0.35 (–0.65, –0.05) 0.024 — — IGF-1, nmol/L –1.60 (–2.89, –0.30) 0.017 –0.92 (–1.50, –0.34) 0.002 0.17 (0.05, 0.30) 0.006 — — Insulin, pmol/L — — –0.21 (–0.38, –0.03) 0.022 0.05 (0.01, 0.09) 0.010 0.02 (–0.01, 0.04) 0.170 Leptin, nmol/L –33.5 (–55.9, –11.2) 0.004 — — — — 4.04 (0.07, 8.01) 0.046 a Model 1 assessed ghrelin and PYY change across time and includes age and age2 as fixed effects and participant ID and intercept as random effects. b Model 2 extended model 1 by testing additional predictors of ghrelin and PYY change using a forward stepwise approach. Additional predictors tested include Tanner stage, height, weight, BMI z score, waist circumference, fat mass, fat-free mass, IGF-1, insulin, leptin, testosterone in boys, and estradiol in girls. Results show only the predictors that significantly contributed to ghrelin or PYY change and were included in the final model. View Large Figure 2. View largeDownload slide Correlation analyses of (a) total ghrelin with height velocity, (b) total ghrelin with weight velocity, (c) total PYY with height velocity, and (d) total PYY with weight velocity. rs, Spearman’s rho. Figure 2. View largeDownload slide Correlation analyses of (a) total ghrelin with height velocity, (b) total ghrelin with weight velocity, (c) total PYY with height velocity, and (d) total PYY with weight velocity. rs, Spearman’s rho. Visual inspection of the PYY plots indicated possible sex-specific trends across age and Tanner stage (Fig. 1). However, these observations were not supported by results of the statistical analyses. Initial mixed modeling did not identify any significant effect of the age-related terms (Table 2, model 1). Similarly, the various functional forms of Tanner stage, including linear (girls: P = 0.55; boys: P = 0.18) and quadratic (girls: P = 0.41; boys: P = 0.24) terms, did not show significant associations with PYY in both sexes (Fig. 1). Serum estradiol and testosterone levels rose significantly across age and Tanner stage in girls and boys, respectively (all P < 0.01; Supplemental Figs. 4 and 5). Insulin and HOMA2-IR increased significantly in boys only (P < 0.05). IGF-1 exhibited significant negative associations with age2 in both sexes (P < 0.01; Supplemental Fig. 4) and Tanner stage2 in girls (P < 0.01; Supplemental Fig. 5), indicating an inverted U-shaped change in IGF-1 over time. Girls showed a significant increase in serum leptin with age (P = 0.03; Supplemental Fig. 4). This is contrasted by leptin levels in boys that remain unchanged with age (Supplemental Fig. 4) and significantly decreased with Tanner stage (Supplemental Fig. 5). Ghrelin and PYY change in relation to physical and biochemical markers of growth Forward stepwise mixed modeling was used to test the contribution of physical and biochemical growth markers to changes in ghrelin and PYY (Table 2, model 2). Results showed that ghrelin in girls was best described using a model that included age, age2, IGF-1, and leptin. Of these, leptin was the strongest predictor (estimate: –33.5; P = 0.004) followed by IGF-1 (estimate: –1.60; P = 0.017), both of which were inversely related to ghrelin in girls. The same statistical procedure applied to ghrelin data in boys yielded a final model that included age, age2, IGF-1, waist circumference, and insulin. Of these, IGF-1 was the strongest predictor (estimate: –0.92; P = 0.002), followed by waist circumference (estimate: –2.30; P = 0.022) and insulin (estimate: –0.21; P = 0.022), all of which were inversely related to ghrelin in boys (Table 2, model 2). Stepwise analyses of PYY data in girls yielded a final model that included age, IGF-1, insulin, and waist circumference (Table 2, model 2). In this model, IGF-1 (estimate: 0.17; P = 0.005) and insulin (estimate: 0.05; P = 0.010) were significant positive predictors, whereas waist circumference (estimate: –0.35; P = 0.024) negatively predicted PYY in girls. Age also showed a marginal positive association with PYY in this model (estimate: 2.07; P = 0.064). Analyses of PYY data in boys revealed a final model that included age, insulin, and leptin. Of these, only leptin was significant and was positively associated with PYY in boys (estimate: 4.04; P = 0.046). Ghrelin and PYY change between weight subgroups To address the secondary aim of the study, we stratified participants into weight subgroups and replicated the plots and predictive models for ghrelin and PYY. Outcomes from initial mixed modeling showed a significant positive age2 effect on ghrelin in healthy-weight girls and boys (Table 3, model 1), consistent with the U-shaped pattern described above (Fig. 3). In contrast, ghrelin levels in Ow/Ob girls exhibited a significant negative association with age2 (estimate: –11.1; P = 0.045), indicative of an inverted U-shaped relationship (Fig. 3), whereas boys with Ow/Ob showed no significant age-related ghrelin change (Table 3, model 1). Tanner stage did not significantly predict ghrelin in any of the weight subgroups (Fig. 3). Unlike the healthy-weight subgroup, all Ow/Ob adolescents reported being well advanced in puberty at baseline (i.e., no participants were at Tanner stage 1). Subsequent mixed-modeling testing additional physical and biochemical predictors revealed a significant negative effect of leptin (estimate: –50.5; P = 0.002) on ghrelin in Ow/Ob girls. In boys, all significant age2, IGF-1, and waist circumference effects on ghrelin identified in the healthy-weight subgroup were not observed in the Ow/Ob adolescents (Table 3, model 2). Table 3. Final Mixed-Effects Models Assessing Ghrelin Change in Relation to Physical and Biochemical Markers of Growth and Development, Analyzed by Sex and Weight Status Total Ghrelin (pmol/L): Girls Total Ghrelin (pmol/L): Boys HW (n = 19) Ow/Ob (n = 7) HW (n = 36) Ow/Ob (n = 18) Estimate (95% CI) P Value Estimate (95% CI) P Value Estimate (95% CI) P Value Estimate (95% CI) P Value Model 1a Age, y –357.6 (–536.9, –178.2) <0.001 318.5 (13.1, 623.8) 0.042 –259.8 (–434.7, –84.9) 0.004 −154.6 (–336.8, 27.7) 0.094 Age2 13.3 (6.75, 19.8) <0.001 –11.1 (–21.9, –0.27) 0.045 9.06 (2.85, 15.3) 0.005 5.26 (–1.15, 11.7) 0.104 Model 2b Age, y –262.9 (–511.0, –14.7) 0.039 282.4 (38.0, 526.7) 0.032 −166.2 (–334.4, 2.08) 0.053 −61.9 (–239.9, 116.2) 0.484 Age2 9.66 (0.32, 19.0) 0.043 –8.85 (–17.4, –0.28) 0.045 6.47 (0.58, 12.4) 0.032 2.23 (–3.96, 8.42) 0.469 Waist, cm — — — — –5.83 (–10.3, –1.35) 0.012 −1.11 (–3.63, 1.41) 0.381 IGF-1, nmol/L −0.39 (–2.04, 1.25) 0.624 −0.46 (–2.43, 1.52) 0.600 –0.18 (–0.30, –0.06) 0.005 −0.08 (–0.17, 0.01) 0.067 Insulin, pmol/L — — — — −0.25 (–0.56, 0.06) 0.112 −0.19 (–0.38, 0.00) 0.050 Leptin, nmol/L −49.0 (–138.1, 40.2) 0.258 –50.5 (–4.75, –1.56) 0.002 — — — — Total Ghrelin (pmol/L): Girls Total Ghrelin (pmol/L): Boys HW (n = 19) Ow/Ob (n = 7) HW (n = 36) Ow/Ob (n = 18) Estimate (95% CI) P Value Estimate (95% CI) P Value Estimate (95% CI) P Value Estimate (95% CI) P Value Model 1a Age, y –357.6 (–536.9, –178.2) <0.001 318.5 (13.1, 623.8) 0.042 –259.8 (–434.7, –84.9) 0.004 −154.6 (–336.8, 27.7) 0.094 Age2 13.3 (6.75, 19.8) <0.001 –11.1 (–21.9, –0.27) 0.045 9.06 (2.85, 15.3) 0.005 5.26 (–1.15, 11.7) 0.104 Model 2b Age, y –262.9 (–511.0, –14.7) 0.039 282.4 (38.0, 526.7) 0.032 −166.2 (–334.4, 2.08) 0.053 −61.9 (–239.9, 116.2) 0.484 Age2 9.66 (0.32, 19.0) 0.043 –8.85 (–17.4, –0.28) 0.045 6.47 (0.58, 12.4) 0.032 2.23 (–3.96, 8.42) 0.469 Waist, cm — — — — –5.83 (–10.3, –1.35) 0.012 −1.11 (–3.63, 1.41) 0.381 IGF-1, nmol/L −0.39 (–2.04, 1.25) 0.624 −0.46 (–2.43, 1.52) 0.600 –0.18 (–0.30, –0.06) 0.005 −0.08 (–0.17, 0.01) 0.067 Insulin, pmol/L — — — — −0.25 (–0.56, 0.06) 0.112 −0.19 (–0.38, 0.00) 0.050 Leptin, nmol/L −49.0 (–138.1, 40.2) 0.258 –50.5 (–4.75, –1.56) 0.002 — — — — Abbreviations: HW, healthy weight. a Model 1 assessed ghrelin change across time and includes age and age2 as fixed effects and participant ID and intercept as random effects. b Model 2 replicated the final model generated from the analysis by sex. Participant ID and intercept were treated as random effects and all other predictor variables as fixed effects. View Large Table 3. Final Mixed-Effects Models Assessing Ghrelin Change in Relation to Physical and Biochemical Markers of Growth and Development, Analyzed by Sex and Weight Status Total Ghrelin (pmol/L): Girls Total Ghrelin (pmol/L): Boys HW (n = 19) Ow/Ob (n = 7) HW (n = 36) Ow/Ob (n = 18) Estimate (95% CI) P Value Estimate (95% CI) P Value Estimate (95% CI) P Value Estimate (95% CI) P Value Model 1a Age, y –357.6 (–536.9, –178.2) <0.001 318.5 (13.1, 623.8) 0.042 –259.8 (–434.7, –84.9) 0.004 −154.6 (–336.8, 27.7) 0.094 Age2 13.3 (6.75, 19.8) <0.001 –11.1 (–21.9, –0.27) 0.045 9.06 (2.85, 15.3) 0.005 5.26 (–1.15, 11.7) 0.104 Model 2b Age, y –262.9 (–511.0, –14.7) 0.039 282.4 (38.0, 526.7) 0.032 −166.2 (–334.4, 2.08) 0.053 −61.9 (–239.9, 116.2) 0.484 Age2 9.66 (0.32, 19.0) 0.043 –8.85 (–17.4, –0.28) 0.045 6.47 (0.58, 12.4) 0.032 2.23 (–3.96, 8.42) 0.469 Waist, cm — — — — –5.83 (–10.3, –1.35) 0.012 −1.11 (–3.63, 1.41) 0.381 IGF-1, nmol/L −0.39 (–2.04, 1.25) 0.624 −0.46 (–2.43, 1.52) 0.600 –0.18 (–0.30, –0.06) 0.005 −0.08 (–0.17, 0.01) 0.067 Insulin, pmol/L — — — — −0.25 (–0.56, 0.06) 0.112 −0.19 (–0.38, 0.00) 0.050 Leptin, nmol/L −49.0 (–138.1, 40.2) 0.258 –50.5 (–4.75, –1.56) 0.002 — — — — Total Ghrelin (pmol/L): Girls Total Ghrelin (pmol/L): Boys HW (n = 19) Ow/Ob (n = 7) HW (n = 36) Ow/Ob (n = 18) Estimate (95% CI) P Value Estimate (95% CI) P Value Estimate (95% CI) P Value Estimate (95% CI) P Value Model 1a Age, y –357.6 (–536.9, –178.2) <0.001 318.5 (13.1, 623.8) 0.042 –259.8 (–434.7, –84.9) 0.004 −154.6 (–336.8, 27.7) 0.094 Age2 13.3 (6.75, 19.8) <0.001 –11.1 (–21.9, –0.27) 0.045 9.06 (2.85, 15.3) 0.005 5.26 (–1.15, 11.7) 0.104 Model 2b Age, y –262.9 (–511.0, –14.7) 0.039 282.4 (38.0, 526.7) 0.032 −166.2 (–334.4, 2.08) 0.053 −61.9 (–239.9, 116.2) 0.484 Age2 9.66 (0.32, 19.0) 0.043 –8.85 (–17.4, –0.28) 0.045 6.47 (0.58, 12.4) 0.032 2.23 (–3.96, 8.42) 0.469 Waist, cm — — — — –5.83 (–10.3, –1.35) 0.012 −1.11 (–3.63, 1.41) 0.381 IGF-1, nmol/L −0.39 (–2.04, 1.25) 0.624 −0.46 (–2.43, 1.52) 0.600 –0.18 (–0.30, –0.06) 0.005 −0.08 (–0.17, 0.01) 0.067 Insulin, pmol/L — — — — −0.25 (–0.56, 0.06) 0.112 −0.19 (–0.38, 0.00) 0.050 Leptin, nmol/L −49.0 (–138.1, 40.2) 0.258 –50.5 (–4.75, –1.56) 0.002 — — — — Abbreviations: HW, healthy weight. a Model 1 assessed ghrelin change across time and includes age and age2 as fixed effects and participant ID and intercept as random effects. b Model 2 replicated the final model generated from the analysis by sex. Participant ID and intercept were treated as random effects and all other predictor variables as fixed effects. View Large Figure 3. View largeDownload slide Mean weight-related differences for (a) total ghrelin with age, (b) total ghrelin with Tanner stage, (c) total PYY with age, and (d) total PYY with Tanner stage. Mixed-model estimates and P values are provided for the quadratic and linear predictive terms, respectively, for the ghrelin and PYY plots. Observations by age in healthy-weight girls: 12 years: n = 8; 13 years: n = 11; 14 years: n = 13; 15 years: n = 7; and 16 years: n = 4. Observations by age in Ow/Ob girls: 12 years: n = 2; 13 years: n = 6; 14 years: n = 8; 15 years: n = 5; and 16 years: n = 1. Observations by age in healthy-weight boys: 12 years: n = 12; 13 years: n = 20; 14 years: n = 24; 15 years: n = 19; and 16 years: n = 13. Observations by age in Ow/Ob boys: 12 years: n = 1; 13 years: n = 11; 14 years: n = 15; 15 years: n = 15; and 16 years: n = 8. HW, healthy-weight; TS, Tanner stage. Figure 3. View largeDownload slide Mean weight-related differences for (a) total ghrelin with age, (b) total ghrelin with Tanner stage, (c) total PYY with age, and (d) total PYY with Tanner stage. Mixed-model estimates and P values are provided for the quadratic and linear predictive terms, respectively, for the ghrelin and PYY plots. Observations by age in healthy-weight girls: 12 years: n = 8; 13 years: n = 11; 14 years: n = 13; 15 years: n = 7; and 16 years: n = 4. Observations by age in Ow/Ob girls: 12 years: n = 2; 13 years: n = 6; 14 years: n = 8; 15 years: n = 5; and 16 years: n = 1. Observations by age in healthy-weight boys: 12 years: n = 12; 13 years: n = 20; 14 years: n = 24; 15 years: n = 19; and 16 years: n = 13. Observations by age in Ow/Ob boys: 12 years: n = 1; 13 years: n = 11; 14 years: n = 15; 15 years: n = 15; and 16 years: n = 8. HW, healthy-weight; TS, Tanner stage. In line with the results across the entire sample, total PYY was not significantly associated with age or Tanner stage in any of the weight subgroups (Fig. 3). PYY in healthy-weight girls showed significant and positive associations with IGF-1 (estimate: 0.02; P = 0.018) and insulin (estimate: 0.04; P = 0.033), of which, only the latter was significant in Ow/Ob girls (Table 4, model 2). No differences were detected for PYY change between healthy-weight and Ow/Ob boys (Table 4). Table 4. Final Mixed-Effects Models Assessing PYY Change in Relation to Physical and Biochemical Markers of Growth and Development, Analyzed by Sex and Weight Status Total PYY (pmol/L): Girls Total PYY (pmol/L): Boys HW (n = 19) Ow/Ob (n = 7) HW (n = 36) Ow/Ob (n = 18) Estimate (95% CI) P Value Estimate (95% CI) P Value Estimate (95% CI) P Value Estimate (95% CI) P Value Model 1a Age, y 2.72 (–0.11, 5.55) 0.059 0.34 (–4.39, 5.10) 0.882 −0.11 (–1.64, 1.43) 0.892 −0.18 (–1.82, 2.17) 0.859 Model 2b Age, y 2.61 (–0.40, 5.62) 0.087 1.23 (–3.14, 5.60) 0.559 −0.24 (–1.81, 1.33) 0.761 0.20 (–1.81, 2.20) 0.846 Waist, cm −0.49 (–1.28, 0.30) 0.216 −0.47 (–0.95,0.01) 0.053 — — — — IGF-1, nmol/L 0.02 (0.00, 0.04) 0.018 0.16 (–0.08, 0.40) 0.159 0.05 (–0.07, 0.17) 0.443 0.81 (–0.02, 0.18) 0.120 Insulin, pmol/L 0.04 (0.00, 0.08) 0.033 0.15 (0.04, 0.25) 0.011 — — — — Total PYY (pmol/L): Girls Total PYY (pmol/L): Boys HW (n = 19) Ow/Ob (n = 7) HW (n = 36) Ow/Ob (n = 18) Estimate (95% CI) P Value Estimate (95% CI) P Value Estimate (95% CI) P Value Estimate (95% CI) P Value Model 1a Age, y 2.72 (–0.11, 5.55) 0.059 0.34 (–4.39, 5.10) 0.882 −0.11 (–1.64, 1.43) 0.892 −0.18 (–1.82, 2.17) 0.859 Model 2b Age, y 2.61 (–0.40, 5.62) 0.087 1.23 (–3.14, 5.60) 0.559 −0.24 (–1.81, 1.33) 0.761 0.20 (–1.81, 2.20) 0.846 Waist, cm −0.49 (–1.28, 0.30) 0.216 −0.47 (–0.95,0.01) 0.053 — — — — IGF-1, nmol/L 0.02 (0.00, 0.04) 0.018 0.16 (–0.08, 0.40) 0.159 0.05 (–0.07, 0.17) 0.443 0.81 (–0.02, 0.18) 0.120 Insulin, pmol/L 0.04 (0.00, 0.08) 0.033 0.15 (0.04, 0.25) 0.011 — — — — Abbreviations: HW, healthy weight. a Model 1 assessed PYY change across time and includes age and age2 as fixed effects and participant ID and intercept as random effects. b Model 2 replicated the final model generated from the analysis by sex. Participant ID and intercept were treated as random effects and all other predictor variables as fixed effects. View Large Table 4. Final Mixed-Effects Models Assessing PYY Change in Relation to Physical and Biochemical Markers of Growth and Development, Analyzed by Sex and Weight Status Total PYY (pmol/L): Girls Total PYY (pmol/L): Boys HW (n = 19) Ow/Ob (n = 7) HW (n = 36) Ow/Ob (n = 18) Estimate (95% CI) P Value Estimate (95% CI) P Value Estimate (95% CI) P Value Estimate (95% CI) P Value Model 1a Age, y 2.72 (–0.11, 5.55) 0.059 0.34 (–4.39, 5.10) 0.882 −0.11 (–1.64, 1.43) 0.892 −0.18 (–1.82, 2.17) 0.859 Model 2b Age, y 2.61 (–0.40, 5.62) 0.087 1.23 (–3.14, 5.60) 0.559 −0.24 (–1.81, 1.33) 0.761 0.20 (–1.81, 2.20) 0.846 Waist, cm −0.49 (–1.28, 0.30) 0.216 −0.47 (–0.95,0.01) 0.053 — — — — IGF-1, nmol/L 0.02 (0.00, 0.04) 0.018 0.16 (–0.08, 0.40) 0.159 0.05 (–0.07, 0.17) 0.443 0.81 (–0.02, 0.18) 0.120 Insulin, pmol/L 0.04 (0.00, 0.08) 0.033 0.15 (0.04, 0.25) 0.011 — — — — Total PYY (pmol/L): Girls Total PYY (pmol/L): Boys HW (n = 19) Ow/Ob (n = 7) HW (n = 36) Ow/Ob (n = 18) Estimate (95% CI) P Value Estimate (95% CI) P Value Estimate (95% CI) P Value Estimate (95% CI) P Value Model 1a Age, y 2.72 (–0.11, 5.55) 0.059 0.34 (–4.39, 5.10) 0.882 −0.11 (–1.64, 1.43) 0.892 −0.18 (–1.82, 2.17) 0.859 Model 2b Age, y 2.61 (–0.40, 5.62) 0.087 1.23 (–3.14, 5.60) 0.559 −0.24 (–1.81, 1.33) 0.761 0.20 (–1.81, 2.20) 0.846 Waist, cm −0.49 (–1.28, 0.30) 0.216 −0.47 (–0.95,0.01) 0.053 — — — — IGF-1, nmol/L 0.02 (0.00, 0.04) 0.018 0.16 (–0.08, 0.40) 0.159 0.05 (–0.07, 0.17) 0.443 0.81 (–0.02, 0.18) 0.120 Insulin, pmol/L 0.04 (0.00, 0.08) 0.033 0.15 (0.04, 0.25) 0.011 — — — — Abbreviations: HW, healthy weight. a Model 1 assessed PYY change across time and includes age and age2 as fixed effects and participant ID and intercept as random effects. b Model 2 replicated the final model generated from the analysis by sex. Participant ID and intercept were treated as random effects and all other predictor variables as fixed effects. View Large Discussion This study investigated pubertal change in fasting circulating levels of total ghrelin and total PYY across 3 years and the relationship of these peptides to anthropometric and biochemical markers of growth and development. Our analyses revealed a U-shaped change in ghrelin that was significantly associated with age2 and IGF-1, but not with Tanner stage. This finding is in contrast to cross-sectional research that documents a linear decrease in fasting ghrelin levels with both age and Tanner stage (16, 17). The strong inverse association between ghrelin and IGF-1 observed in our current study is in line with cross-sectional data (17). However, the absence of a significant Tanner stage effect on ghrelin, together with the nonsignificant testosterone and estradiol associations, suggests that adolescent ghrelin levels may be more strongly linked to markers of somatic growth than sexual maturation. Absence of significant estradiol and testosterone associations support the concept that gonadal hormones indirectly affect growth via GH and IGF-1 production (30). Also, Tanner stages are a distal measure of gonadal hormone change and are subject to measurement bias, especially when data are self-reported (31), as is the case for this study. Importantly, analyses of age at ghrelin nadir vs age at peak height velocity showed no differences in the timing of these two events, and significant negative correlations were identified between ghrelin and annual height and weight velocity. Taken together, these two lines of evidence strongly support a drop in circulating total ghrelin to its lowest levels during peak pubertal growth and allow the proposition that low ghrelin concentrations signal adequate nutritional status to support rapid somatic growth and development of reproductive capacity (32). Thus, we propose that the strong appetites of pubertal adolescents cannot be solely attributed to ghrelin and its orexigenic properties. Although a potential increase in sensitivity to the hunger-promoting effects of ghrelin cannot be ruled out, as postulated in one study (17), our data lend greater support to the role of ghrelin in regulating long-term energy balance and pubertal development. Some studies have reported that circulating ghrelin, together with leptin and insulin, operate as opposing energy balance signals in the control of pubertal onset and progression (14, 33). Results from our mixed models suggest this metabolic control of puberty may be sex specific, as ghrelin was strongly associated with leptin but not insulin in girls, and vice versa in boys. Although energy balance was not measured in this study, there is increasing recognition of ghrelin as a chronic starvation signal, as evidenced by a well-documented inverse association between fasting ghrelin and BMI (7) and the low and high ghrelin levels reported in over- and undernutrition states, respectively (7). Supportive of ghrelin’s signaling of undernutrition are its proposed inhibitory influences on fertility and maturation (32, 34). Such inhibitory influences may be interpreted as an adaptive response to suboptimal nutrition states for reproduction. In rodents, repeated ghrelin infusions suppressed gonadotropin release and delayed pubertal onset (35). Our current data show an early to midpubertal fall in ghrelin, which would thus be permissive to initiation and progression of puberty. This is associated with the observed rise in insulin and IGF-1 promoting rapid physical growth. Negative feedback from GH may further suppress ghrelin levels during this period (36). This study found no significant relationship for total PYY to age, Tanner stage, or physical markers of growth in either sex. This outcome is consistent with research showing similar fasting PYY levels in prepubertal vs pubertal youth (20), but differs from the findings of a cross-sectional study by Lloyd et al. (13). In the Lloyd study of similar sample size, midpubertal teens were reported to have significantly lower fasting PYY compared with those in early or late puberty (13). PYY levels correlated inversely with GH, and GH was postulated as the driver of the midpubertal PYY nadir and subsequent increases in appetite and gonadotropins. In contrast, our current work showed a small but significant positive relationship between PYY, IGF-1, and insulin in girls. This result suggests a slight increase in satiety signaling with pubertal growth, which was not anticipated. Comparison of our results to those from the Lloyd study showed similar fasting PYY levels at pre/peripuberty (42.4 vs 30.3 pmol/L in Lloyd et al.) (13). The main difference was our relatively small magnitude of PYY change, which potentially reflects the longitudinal nature of our data. As this study observed a minimal change in fasting PYY levels across puberty, perhaps a blunted satiety response after food intake may explain the increased appetite typically observed in rapidly growing adolescents (37). Comparison of ghrelin and PYY alterations by weight subgroup revealed that, unlike healthy-weight adolescents, those with Ow/Ob did not exhibit a U-shaped change in ghrelin with age. This finding is in contrast to a previous cross-sectional study in which ghrelin levels dropped equally among healthy-weight and Ow/Ob youth (38). Loss of the U-shaped trend may represent an adaptive response to chronic energy surplus in the Ow/Ob adolescents or alternatively, it may reflect true dysregulation of energy balance signals with implications for long-term weight and metabolic risk trajectories. The small size of the Ow/Ob subgroups, who also reported being well advanced in puberty at the start of the study, meant that we are unable to draw any firm conclusions about weight-related differences in adolescent ghrelin and PYY levels. Longitudinal data from an Ow/Ob cohort who are less developmentally advanced would help clarify this apparent dysregulation. The longitudinal study design in a community sample and comprehensive assessment of anthropometric, body composition, and biochemical data strengthen the unique findings of this research. The lack of pre- and postpuberty data, the small sample size for Ow/Ob subgroup analyses, and assessment of Tanner stage by self-report are limitations. The plasma sampling procedures precluded measurement of ghrelin and PYY in its different forms. Also, with annual anthropometry, it is not possible to ascertain how weight fluctuation and active attempts to gain/lose weight may influence ghrelin and PYY levels. Conclusion This study showed, in a healthy longitudinal adolescent cohort, a negative association for fasting total ghrelin with annual height and weight velocity. This association yielded a U-shaped relationship with age, which was inverse to the changes observed for IGF-1. The strongest predictor of the ghrelin-age relationship was leptin in girls and IGF-1 in boys. Adolescents with Ow/Ob showed an apparent loss of the U-shaped ghrelin-age relationship, but this finding may be attributed to their greater level of maturity. Contrary to cross-sectional data, a direct association for ghrelin to Tanner stage or gonadal hormones was not demonstrable. The current study did not identify any significant relationship for total PYY to age, Tanner stage, or physical growth. Insulin and IGF-1 in girls and leptin in boys were significant positive predictors of PYY. Taken together, results of this study suggest that, during puberty, the role of ghrelin is directed to ensuring appropriate energy balance to accommodate the growth spurt, given the identified relationships with IGF-1 and leptin. The relationship of gonadal hormones to both ghrelin and PYY production remains unclear. Abbreviations: Abbreviations: BMI body mass index HOMA2-IR updated homeostasis model assessment-insulin resistance Ow/Ob overweight and obesity PYY peptide YY Acknowledgments We thank Dr. Leonel Prado-Lourenço for his guidance on laboratory setup and obtaining university radiation project clearance, Dr. Fiona Atkinson for her assistance with glucose analysis, Mazen Amatoury for his contribution in managing the ARCHER study biobank, and Dr. Radhika Seimon for her assistance with reviewing the revised manuscript. Financial Support: This work was supported by a joint Sydney Medical School and Balnaves Foundation Kick Start Grant (to H.L.C.), a Sydney Medical School Lifespan Seed Grant (to H.L.C.), the National Health and Medical Research Council of Australia (to K.S.) and the Thyne Reid Foundation (to K.S.). Clinical Trial Information: Australian New Zealand ClinicalTrials.gov no. ACTRN12617000964314 (registered 5 July 2017). Author Contributions: H.L.C., A.S., K.P., G.L., C.H., and K.S. designed research, H.L.C., A.S., M.S., and K.P. conducted research, H.L.C. and A.S. provided essential laboratory advice and materials, H.L.C., F.G., and M.S. analyzed data, H.L.C., A.S., F.G., and K.S. interpreted data and wrote the paper, and H.L.C. had primary responsibility for final content. All authors read and approved the final manuscript. Disclosure Summary: A.S. is the author of The Don’t Go Hungry Diet and Don’t Go Hungry For Life. She has also received payment from Eli Lilly, the Pharmacy Guild of Australia, Novo Nordisk, the Dietitians Association of Australia, Shoalhaven Family Medical Centres, and the Pharmaceutical Society of Australia for presentation at conferences and has served on the Nestlé Health Science Optifast VLCD advisory board since 2016. The remaining authors have nothing to disclose. References 1. Rogol AD , Roemmich JN , Clark PA . Growth at puberty . J Adolesc Health . 2002 ; 31 ( 6 , Suppl ) 192 – 200 . Google Scholar CrossRef Search ADS PubMed 2. Kerrigan JR , Rogol AD . The impact of gonadal steroid hormone action on growth hormone secretion during childhood and adolescence . Endocr Rev . 1992 ; 13 ( 2 ): 281 – 298 . Google Scholar PubMed 3. Jeffery AN , Metcalf BS , Hosking J , Streeter AJ , Voss LD , Wilkin TJ . Age before stage: insulin resistance rises before the onset of puberty: a 9-year longitudinal study (EarlyBird 26) . Diabetes Care . 2012 ; 35 ( 3 ): 536 – 541 . Google Scholar CrossRef Search ADS PubMed 4. Cheng HL , Amatoury M , Steinbeck K . Energy expenditure and intake during puberty in healthy nonobese adolescents: a systematic review . Am J Clin Nutr . 2016 ; 104 ( 4 ): 1061 – 1074 . Google Scholar CrossRef Search ADS PubMed 5. Mostazir M , Jeffery A , Hosking J , Metcalf B , Voss L , Wilkin T . Evidence for energy conservation during pubertal growth. A 10-year longitudinal study (EarlyBird 71) . Int J Obes . 2016 ; 40 ( 11 ): 1619 – 1626 . Google Scholar CrossRef Search ADS 6. Sun M , Gower BA , Bartolucci AA , Hunter GR , Figueroa-Colon R , Goran MI . A longitudinal study of resting energy expenditure relative to body composition during puberty in African American and white children . Am J Clin Nutr . 2001 ; 73 ( 2 ): 308 – 315 . Google Scholar CrossRef Search ADS PubMed 7. Suzuki K , Jayasena CN , Bloom SR . The gut hormones in appetite regulation . J Obes . 2011 ; 2011 : 528401 . Google Scholar CrossRef Search ADS PubMed 8. Pradhan G , Samson SL , Sun Y . Ghrelin: much more than a hunger hormone . Curr Opin Clin Nutr Metab Care . 2013 ; 16 ( 6 ): 619 – 624 . Google Scholar CrossRef Search ADS PubMed 9. Tena-Sempere M . Interaction between energy homeostasis and reproduction: central effects of leptin and ghrelin on the reproductive axis . Horm Metab Res . 2013 ; 45 ( 13 ): 919 – 927 . Google Scholar CrossRef Search ADS PubMed 10. Sumithran P , Prendergast LA , Delbridge E , Purcell K , Shulkes A , Kriketos A , Proietto J . Long-term persistence of hormonal adaptations to weight loss . N Engl J Med . 2011 ; 365 ( 17 ): 1597 – 1604 . Google Scholar CrossRef Search ADS PubMed 11. Guo Y , Ma L , Enriori PJ , Koska J , Franks PW , Brookshire T , Cowley MA , Salbe AD , Delparigi A , Tataranni PA . Physiological evidence for the involvement of peptide YY in the regulation of energy homeostasis in humans . Obesity (Silver Spring) . 2006 ; 14 ( 9 ): 1562 – 1570 . Google Scholar CrossRef Search ADS PubMed 12. Tovar SA , Seoane LM , Caminos JE , Nogueiras R , Casanueva FF , Diéguez C . Regulation of peptide YY levels by age, hormonal, and nutritional status . Obes Res . 2004 ; 12 ( 12 ): 1944 – 1950 . Google Scholar CrossRef Search ADS PubMed 13. Lloyd B , Ravi P , Mendes N , Klibanski A , Misra M . Peptide YY levels across pubertal stages and associations with growth hormone . J Clin Endocrinol Metab . 2010 ; 95 ( 6 ): 2957 – 2962 . Google Scholar CrossRef Search ADS PubMed 14. Horner K , Lee S . Appetite-related peptides in childhood and adolescence: role of ghrelin, PYY, and GLP-1 . Appl Physiol Nutr Metab . 2015 ; 40 ( 11 ): 1089 – 1099 . Google Scholar CrossRef Search ADS PubMed 15. Gordon-Larsen P , Adair LS , Nelson MC , Popkin BM . Five-year obesity incidence in the transition period between adolescence and adulthood: the National Longitudinal Study of Adolescent Health . Am J Clin Nutr . 2004 ; 80 ( 3 ): 569 – 575 . Google Scholar PubMed 16. Soriano-Guillén L , Barrios V , Chowen JA , Sánchez I , Vila S , Quero J , Argente J . Ghrelin levels from fetal life through early adulthood: relationship with endocrine and metabolic and anthropometric measures . J Pediatr . 2004 ; 144 ( 1 ): 30 – 35 . Google Scholar CrossRef Search ADS PubMed 17. Whatmore AJ , Hall CM , Jones J , Westwood M , Clayton PE . Ghrelin concentrations in healthy children and adolescents . Clin Endocrinol (Oxf) . 2003 ; 59 ( 5 ): 649 – 654 . Google Scholar CrossRef Search ADS PubMed 18. Pomerants T , Tillmann V , Jürimäe J , Jürimäe T . Relationship between ghrelin and anthropometrical, body composition parameters and testosterone levels in boys at different stages of puberty . J Endocrinol Invest . 2006 ; 29 ( 11 ): 962 – 967 . Google Scholar CrossRef Search ADS PubMed 19. Võsoberg K , Tillmann V , Tamm A-L , Jürimäe T , Saar M , Maasalu K , Neissaar I , Lätt E , Jürimäe J . Adipocytokine and ghrelin levels in relation to body composition in rhythmic gymnasts entering into puberty: a 3-year follow-up study . Pediatr Exerc Sci . 2014 ; 26 ( 4 ): 477 – 484 . Google Scholar CrossRef Search ADS PubMed 20. Roth CL , Enriori PJ , Harz K , Woelfle J , Cowley MA , Reinehr T . Peptide YY is a regulator of energy homeostasis in obese children before and after weight loss . J Clin Endocrinol Metab . 2005 ; 90 ( 12 ): 6386 – 6391 . Google Scholar CrossRef Search ADS PubMed 21. Amon K , Paxton K , Klineberg E , Riley L , Hazell P , Skinner SR , Hawke C , Steinbeck K . Recruiting a young adolescent rural cohort: costs and lessons learnt . Adv Pediatr Res . 2016 ; 3 : 1 – 9 . Google Scholar CrossRef Search ADS 22. Steinbeck K , Hazell P , Cumming RG , Skinner SR , Ivers R , Booy R , Fulcher G , Handelsman DJ , Martin AJ , Morgan G , Starling J , Bauman A , Rawsthorne ML , Bennett DL , Chow CM , Lam MK , Kelly P , Brown NJ , Paxton K , Hawke C . The study design and methodology for the ARCHER study--adolescent rural cohort study of hormones, health, education, environments and relationships . BMC Pediatr . 2012 ; 12 ( 1 ): 143 . Google Scholar CrossRef Search ADS PubMed 23. Chula de Castro JA , Lima TR , Silva DAS . Body composition estimation in children and adolescents by bioelectrical impedance analysis: a systematic review . J Bodyw Mov Ther . 2018 ; 22 ( 1 ): 134 – 146 . Google Scholar CrossRef Search ADS PubMed 24. Cole TJ , Lobstein T . Extended international (IOTF) body mass index cut-offs for thinness, overweight and obesity . Pediatr Obes . 2012 ; 7 ( 4 ): 284 – 294 . Google Scholar CrossRef Search ADS PubMed 25. Garnett SP , Baur LA , Cowell CT . Waist-to-height ratio: a simple option for determining excess central adiposity in young people . Int J Obes . 2008 ; 32 ( 6 ): 1028 – 1030 . Google Scholar CrossRef Search ADS 26. Marshall WA , Tanner JM . Variations in the pattern of pubertal changes in boys . Arch Dis Child . 1970 ; 45 ( 239 ): 13 – 23 . Google Scholar CrossRef Search ADS PubMed 27. Marshall WA , Tanner JM . Variations in pattern of pubertal changes in girls . Arch Dis Child . 1969 ; 44 ( 235 ): 291 – 303 . Google Scholar CrossRef Search ADS PubMed 28. Harwood DT , Handelsman DJ . Development and validation of a sensitive liquid chromatography-tandem mass spectrometry assay to simultaneously measure androgens and estrogens in serum without derivatization . Clin Chim Acta . 2009 ; 409 ( 1-2 ): 78 – 84 . Google Scholar CrossRef Search ADS PubMed 29. Levy JC , Matthews DR , Hermans MP . Correct homeostasis model assessment (HOMA) evaluation uses the computer program . Diabetes Care . 1998 ; 21 ( 12 ): 2191 – 2192 . Google Scholar CrossRef Search ADS PubMed 30. Perry RJ , Farquharson C , Ahmed SF . The role of sex steroids in controlling pubertal growth . Clin Endocrinol (Oxf) . 2008 ; 68 ( 1 ): 4 – 15 . Google Scholar CrossRef Search ADS PubMed 31. Huang B , Hillman J , Biro FM , Ding L , Dorn LD , Susman EJ . Correspondence between gonadal steroid hormone concentrations and secondary sexual characteristics assessed by clinicians, adolescents, and parents . J Res Adolesc . 2012 ; 22 ( 2 ): 381 – 391 . Google Scholar CrossRef Search ADS PubMed 32. Tena-Sempere M . Ghrelin as a pleotrophic modulator of gonadal function and reproduction . Nat Clin Pract Endocrinol Metab . 2008 ; 4 ( 12 ): 666 – 674 . Google Scholar CrossRef Search ADS PubMed 33. Abreu AP , Kaiser UB . Pubertal development and regulation . Lancet Diabetes Endocrinol . 2016 ; 4 ( 3 ): 254 – 264 . Google Scholar CrossRef Search ADS PubMed 34. Méquinion M , Langlet F , Zgheib S , Dickson S , Dehouck B , Chauveau C , Viltart O . Ghrelin: central and peripheral implications in anorexia nervosa . Front Endocrinol (Lausanne) . 2013 ; 4 : 15 . Google Scholar PubMed 35. Fernández-Fernández R , Navarro VM , Barreiro ML , Vigo EM , Tovar S , Sirotkin AV , Casanueva FF , Aguilar E , Dieguez C , Pinilla L , Tena-Sempere M . Effects of chronic hyperghrelinemia on puberty onset and pregnancy outcome in the rat . Endocrinology . 2005 ; 146 ( 7 ): 3018 – 3025 . Google Scholar CrossRef Search ADS PubMed 36. Edén Engström B , Burman P , Holdstock C , Karlsson FA ; Engström BEn . Effects of growth hormone (GH) on ghrelin, leptin, and adiponectin in GH-deficient patients . J Clin Endocrinol Metab . 2003 ; 88 ( 11 ): 5193 – 5198 . Google Scholar CrossRef Search ADS PubMed 37. Patel BP , Anderson GH , Vien S , Bellissimo N , McCrindle BW , Hamilton JK . Obesity, sex and pubertal status affect appetite hormone responses to a mixed glucose and whey protein drink in adolescents . Clin Endocrinol (Oxf) . 2014 ; 81 ( 1 ): 63 – 70 . Google Scholar CrossRef Search ADS PubMed 38. Bellone S , Prodam F , Savastio S , De Rienzo F , Demarchi I , Trovato L , Petri A , Rapa A , Aimaretti G , Bona G . Acylated and unacylated ghrelin levels in normal weight and obese children: influence of puberty and relationship with insulin, leptin and adiponectin levels . J Endocrinol Invest . 2012 ; 35 ( 2 ): 191 – 197 . Google Scholar PubMed Copyright © 2018 Endocrine Society

Journal of Clinical Endocrinology and Metabolism – Oxford University Press

**Published: ** Aug 1, 2018

Loading...

personal research library

It’s your single place to instantly

**discover** and **read** the research

that matters to you.

Enjoy **affordable access** to

over 18 million articles from more than

**15,000 peer-reviewed journals**.

All for just $49/month

Query the DeepDyve database, plus search all of PubMed and Google Scholar seamlessly

Save any article or search result from DeepDyve, PubMed, and Google Scholar... all in one place.

Get unlimited, online access to over 18 million full-text articles from more than 15,000 scientific journals.

Read from thousands of the leading scholarly journals from *SpringerNature*, *Elsevier*, *Wiley-Blackwell*, *Oxford University Press* and more.

All the latest content is available, no embargo periods.

## “Hi guys, I cannot tell you how much I love this resource. Incredible. I really believe you've hit the nail on the head with this site in regards to solving the research-purchase issue.”

Daniel C.

## “Whoa! It’s like Spotify but for academic articles.”

@Phil_Robichaud

## “I must say, @deepdyve is a fabulous solution to the independent researcher's problem of #access to #information.”

@deepthiw

## “My last article couldn't be possible without the platform @deepdyve that makes journal papers cheaper.”

@JoseServera

DeepDyve ## Freelancer | DeepDyve ## Pro | |
---|---|---|

Price | FREE | $49/month |

Save searches from | ||

Create lists to | ||

Export lists, citations | ||

Read DeepDyve articles | Abstract access only | Unlimited access to over |

20 pages / month | ||

PDF Discount | 20% off | |

Read and print from thousands of top scholarly journals.

System error. Please try again!

or

By signing up, you agree to DeepDyve’s Terms of Service and Privacy Policy.

Already have an account? Log in

Bookmark this article. You can see your Bookmarks on your DeepDyve Library.

To save an article, **log in** first, or **sign up** for a DeepDyve account if you don’t already have one.

All DeepDyve websites use cookies to improve your online experience. They were placed on your computer when you launched this website. You can change your cookie settings through your browser.

ok to continue