Pediatric reference values for glucagon-like peptide-1 (GLP-1), glucose-dependent insulinotropic polypeptide (GIP), and glucagon in children and adolescents

Abstract

Background

Incretin hormones, including glucose-dependent insulinotropic polypeptide (GIP) and glucagon-like peptide-1 (GLP-1), along with glucagon, are regulators of appetite, energy balance, and fat metabolism. Despite their recognized functions, reference values for these hormones in children have not previously been established. This study aimed to establish age- and sex-specific reference percentiles and z-scores for GLP-1, GIP, and glucagon in children and to examine their associations with insulin resistance, lipid status, and metabolic syndrome.

Methods

Data were derived from the cohort of the study Childhood Health, Activity, and Motor Performance School Study Denmark (CHAMPS study DK) including 485 school children without metabolic syndrome.

Results

Age was positively associated with GLP-1 levels in both sexes, while GIP levels showed an inverse relationship with age in girls. Insulin sensitivity, as measured by Homeostatic Model Assessment 2 for Insulin Resistance (HOMA2-IR), was positively associated with GLP-1 z-scores in boys and glucagon z-scores in girls.

Conclusion

Age, sex and insulin sensitivity are important determinants of incretin and glucagon hormone levels in children. The reference percentiles and z-scores generated in this study provide a framework for future studies in pediatric metabolic health and disease.

Impact

• This study provides age- and sex-specific reference values for glucagon-like peptide-1 (GLP-1), glucose-dependent insulinotropic polypeptide (GIP), and glucagon in children and adolescents.

• This is the first publication presenting percentiles and z-scores for GLP-1 and GIP in children which is of value for future pediatric comparisons.

• Knowledge about normal physiology is of major importance for future research in the upcoming area of GLP-1, GIP and glucagon as potential biomarkers for metabolic disorders.

Introduction

In healthy individuals, the maintenance of plasma glucose homeostasis is obtained by finely tuned interactions of hormones, metabolic processes, energy uptake and expenditure.¹ Alterations in this fine balance can lead to obesity, prediabetes and other metabolic diseases, which have become an increasing concern in children and adolescents.² Glucagon and incretin hormones have potential as targets in the treatment of obesity and metabolic disorders, as they directly influence the fundamental processes of appetite regulation, energy homeostasis, and fat metabolism.^3,4

Clinical interpretation of health versus disease depends severely on the availability of appropriate reference intervals. Reference material have been established in healthy children and adolescents for glucose, insulin, insulin resistance,⁵ bone turnover markers,⁶ adipokines,⁷ lipids⁸ and multiple other metabolites and hormones.⁹ However, reference values on glucose co-regulatory hormones, like incretins and glucagon, have to our knowledge not previously been published for children.

The best characterized incretin hormones are glucose-dependent insulinotropic polypeptide (GIP) and glucagon-like peptide 1 (GLP-1). Both GIP and GLP-1 are primarily secreted by enteroendocrine cells within the intestines after nutrient ingestion.¹⁰ Properties of GIP include the ability to stimulate insulin secretion during postprandial conditions and glucagon secretion in both euglycemic and hypoglycemic conditions to restore euglycemia.¹¹ GIP is also thought to enhance storage of triglycerides in adipose tissue and contribute to bone remodeling.^3,12,13

In addition to stimulating insulin secretion, GLP-1 inhibits glucagon secretion, and thereby decreases plasma glucose levels. GLP-1 also reduces appetite by increasing satiety and prolonging gastric emptying which, in combination, results in weight-loss.¹²

It is estimated that more than 20% of Danish children and adolescents are overweight and 4–5% are obese.^14,15 Obesity can lead to metabolic syndrome that can further increase the risk of heart disease, diabetes and other health conditions. The mechanisms involved in the development of prediabetes is thought to be a combination of increased adipose tissue resulting in low grade inflammation, insulin resistance and dysfunction of the beta cells in the pancreas.² This metabolic dysfunction is still not fully understood but has previously been proposed to involve incretins and glucagon.¹² The primary roles of glucagon in healthy individuals are glycogenolysis, gluconeogenesis, ketogenesis and lipolysis to promote glucose and alternative energy sources to be released into circulation.⁴ Moreover, glucagon has an inhibitory effect on insulin secretion from the pancreatic beta cells. Studies in adults suggest that fasting hyperglucagonemia can be an early sign of dysmetabolism in obesity.¹⁶ As incretins play a central role in the entero-insular axis, subtle dysregulations could be early biomarkers of disease.¹⁶ However, as there might be age dependent variation in normal incretin physiology in children, age- and sex specific reference intervals are essential for distinguishing between normal developmental physiology of the incretins - and pathology - e.g., metabolic syndrome.

GLP-1 receptor agonists are emerging as a safe and efficient treatment option for children with obesity or metabolic syndrome,^17,18 and some studies suggest that baseline levels of incretins might even be able to predict individual responses,¹⁹ further understating a need to qualify reference intervals.

The aim of this study was to determine age- and sex- specific reference percentile values and z-scores of GLP-1, GIP and glucagon in children. We also investigate the associations between insulin resistance, pubertal stage, lipid status and metabolic syndrome to GLP-1, GIP and glucagon z-scores.

Materials and methods

Participants

This cohort is based on data from The Childhood Health, Activity, and Motor Performance School Study Denmark (CHAMPS-Study DK or CHAMPS)–a prospective controlled intervention study evaluating health outcomes based on increased physical activity, conducted in the Danish municipality of Svendborg from 2008 to 2015. Details about the CHAMPS-Study DK have previously been published.^20,21 A total of ten primary schools were included, six of whom were “sports schools” (intervention) and four matched control schools. The CHAMPS study included 3 visits: a baseline visit prior to intervention in 2008 and two follow-up visits in 2010 and 2015. The current study involves a representative sub-cohort (including baseline visit and follow-up visits from control schools). Inclusion to the current reduced cohort required (i) data available on anthropometrics (age, height, weight), (ii) fasting blood samples and (iii) available measurements on either GLP-1, GIP or glucagon. This current cohort resulted in 506 individuals in total. A sub-cohort of the current cohort comprised 464 children and adolescents after excluding children with metabolic syndrome. The 464 children were used at the ‘reference group’ when generating age- and sex- specific percentiles and z-scores.

Anthropometrics

Throughout the anthropometric measurements’ children wore light clothes and no shoes. Using a stadiometer, height was measured to the nearest centimeter and weight was measured to the nearest 100 g using a calibrated portable electronic weight. Waist circumference (WC) was measured twice at umbilical level to the nearest 0.5 cm using a measuring tape, if the difference between these two measurements exceeded 1 cm a third measurement was conducted. The mean of the two nearest WC measurements was then calculated. Using the formula: weight divided by height squared (kg/m²), Body mass index (BMI) was calculated. By means of Danish references for anthropometric measurements, WC and BMI were transformed into z-scores.²²

Dual energy X-ray absorptiometry

All dual energy X-ray absorptiometry (DXA) scans were performed using GE Lunar Prodigy (GE Medical Systems, Madison, WI), ENCORE software (version 12.3, Prodigy; Lunar Corp, Madison, WI). The scans were performed by two operators, but the analyses were done by one person. We calculated the percentage of total body fat mass as follows: Total body fat mass percentage = (total body fat mass/total body mass x 100).

Pubertal stage

Based on drawings and text representing Tanner stages for breast 1–5 (female) and pubic hair 1–5 (male), the pubertal stage was self-assessed. As self-assessment is only viable in distinguishing puberty vs. no puberty,²³ we chose to dichotomize puberty into prepubertal (Tanner stage 1/T1) and pubertal (Tanner stage 2-5/T2-T5) in this study.

Insulin sensitivity

We used two surrogate measures to estimate insulin sensitivity: the Homeostatic Model Assessment for Insulin Resistance version 1 (HOMA-IR) and version 2 (HOMA2-IR). Both surrogate measures are based on fasting levels of plasma insulin and glucose. HOMA-IR was calculated as insulin mU/L × glucose mM/22.5, and HOMA2-IR was calculated by a computerized algorithm. Further details regarding data collection and procedures has been described in a previous study.²⁰

Metabolic syndrome

The criteria for describing metabolic syndrome (MetS) in this paper was defined by Ahrens et al.²⁴ MetS was defined by having three or more of the following: blood pressure (>90th percentile systolic or diastolic), waist circumference (>90th percentile), lipid profile (triglycerides (>90th percentile and/or High Density Lipoprotein (HDL) <10th percentile) or HOMA-IR >90th percentile or fasting glucose >90th percentile. The children with MetS were excluded from the ‘reference’ group.

Blood chemistry and biomarkers

Blood samples were collected between 08:00 and 10:00 following an overnight fast. Samples were processed within four hours and stored at −80 °C until further analysis. Glucose concentrations were determined using an enzymatic, colorimetric method on Roche/Hitachi Cobas c systems. Insulin levels were measured using a solid-phase enzyme-labeled chemiluminescent immunometric assay.

Incretin hormones, GIP (total) GLP-1 (total) and glucagon, were quantified using Mesoscale Discovery (MSD) assays on the MESO QuickPlex SQ 120 instrument (Meso Scale Diagnostics). These assays were based on analyte-specific antibody pairs and the capture antibody binds to up to 10 spots in a 96-well plate. The detection antibody then generates a light signal proportional to the amount of bound analyte. The light signals were subsequently converted to concentrations using standard curves specific to each assay plate.

Each Mesoscale analysis plate included duplicates of eight standards and a pooled control sample. The pool data was used to assess intra- and inter-plate variation. Samples were analyzed with a single determination. Prior to analysis, samples were thawed on ice, centrifuged, and diluted 1:1 with a working solution containing valine-pyrrolidide (Val-pyr) for dipeptidyl-peptidase-4 (DPP4) inhibition.

Sample pipetting was done both manually and using a BRAVO robotic system, while washing steps were done with a 96-channel plate washer (Biotek 405TS). Total GIP concentrations were measured using U-PLEX antibodies (B215S) combined with nine additional analytes on a U-PLEX 10-assay plate (N05235A-1), following the manufacturer’s protocol. Of the 506 samples analyzed, 55% were within the quantification range of the GIP assay, with an additional 35% within the detection range. Intra-plate variation for the pooled control was 14.7% and inter-plate variation was 25.6%.

Glucagon and total GLP-1 concentrations were measured using V-PLEX assays (Glucagon: K150U5D, GLP-1 (total): K1503PD). The same pre-analytical procedures as for the GIP assay were used, with all samples undergoing two freeze-thaw cycles before measurement. All 506 samples which were analyzed for total GLP-1 and Glucagon were within the quantification ranges of the MSD assays. Glucagon intra-plate variation was 6.5%, and inter-plate variation was 16.1%. For total GLP-1, the intra- and inter-plate variations were 6.6% and 12.1%, respectively.

Statistics

All statistical analyses were performed using R Studio version 4.1.3, and p < 0.05 were considered statistically significant. Data were analyzed as cross-sectional data.

To provide an overview, all data are presented as median [IQR] or standard deviations (SD) for continuous data, and as number (%) for categorical data. To estimate the differences between boys and girls Mann–Whitney U test was used on non-normally distributed data and Student’s T test on normally distributed data. Normal distribution of data was assessed using histograms, qq-plots and Shapiro Wilk tests for normality. Based on the LMS-method, z-scores of BMI and WC were calculated.²⁵ Age- and sex-specific percentile values of incretin hormones and glucagon were calculated using the General Additive Model for Location, Scale and Shape (GAMLSS) with the Box-Cox Cole and Green (BCCG) distribution and z-score values were calculated thereafter. Extreme outliers were removed using the Tukey fences (IQR) method with k set to three. The fitted models used did not depend on normally distributed data, hence plasma concentrations did not need to be log-transformed in this regard.

Linear regression analysis, adjusted for BMI z-score, was used to evaluate associations between z-scores (incretins and glucagon) and selected variables. The interaction terms between sex and the selected variables were used to determine whether the association differed significantly dependent on sex. For the multiple linear regression analysis, the variables were standardized using the ‘scale function’ in R resulting in standardized beta coefficients outputs.

Ethical approval

Before initiation of the project, the CHAMPS-Study DK was approved by the Ethics committee (project ID: S20080047) and registered in the Danish Data Protection Agency as required by Danish law J.nr. 2008-41-2240.

Results

Study population

A summary of the study population without children with metabolic syndrome is presented in Table 1. The median age of the study population was 10.80 years [9.56–12.61]. Boys and girls were comparable on anthropometric measurements except for waist circumference z-score where the waist z-score was marginally larger in girls (p = 0.051). The lipid profile differed between boys and girls, with girls having higher median levels of Low-Density Protein cholesterol (LDL cholesterol) (p < 0.001), triglycerides (p < 0.001), and total cholesterol (p < 0.001). Girls also had significantly higher total fat mass measured by DXA compared to boys (p < 0.001) and the insulin resistance surrogate markers HOMA-IR and HOMA2-IR score were also higher in girls compared to boys (p < 0.001). More boys were self-assessed as being pubertal (68%) compared to 52% of the girls (p = 0.001).

Table 1.

Baseline characteristics and hormone levels of the study population

	Total	Girls	Boys	P-value
Anthropometrics
n (%)	485	218 (45)	267 (55)
Age (years)	10.80 [9.56, 12.61]	10.71 [9.73, 12.44]	10.84 [9.46, 12.78]	0.870
Height (cm)	146.0 [137.9,157.0]	146[137.2, 156.4]	146.5 [138.5, 159.25]	0.226
Weight (kg)	36.70 [31.58, 45.90]	36.00 [31.18, 44.17]	36.70 [31.45, 45.80]	0.431
BMI (kg/m2)	17.20 [15.70, 18.90]	17.30 [15.70, 19.00]	17.00 [15.60, 18.60]	0.498
BMI z-score	−0.08 (1.05)	−0.01 (1.01)	−0.16 (1.08)	0.076
Waist circumf. (cm)	63.50 [59.0, 70.00]	63.0 (58.4, 68.6)	64.0 [59.0, 70.0]	0.460
Waist circumf. z-score	0.14 (1.08)	0.26 (1.09)	0.04 (1.06)	0.051
Total fat mass % (DXA)	20.28 [15.10, 27.01]	24.21 [19.84, 29.85]	16.25 [11.97, 21.98]	<0.001
Puberty
Pubertal (T2-5), n (%)	273 (61.3)	110 (53.9)	163 (67.6)	0.001
Lipids
HDL (mg/dL)	1.58 [1.34, 1.89]	1.55 [1.37, 1.84]	1.60 [1.34, 1.91]	0.565
LDL (mg/dL)	2.25 [1.89, 2.69]	2.43 [2.07, 2.79]	2.17 [1.78, 2.56]	<0.001
Triglycerides (mg/dL)	0.58 [0.47, 0.76]	0.63 [0.52, 0.82]	0.54 [0.45, 0.70]	<0.001
Total cholesterol (mg/dL)	4.24 [3.75, 4.65]	4.37 [3.94, 4.75]	4.14 [3.63, 4.55]	<0.001
Biomarkers
GLP-1 (pmol/L)	4.64 [3.81, 5.65]	4.56 [3.65, 5.61]	4.73 [3.89, 5.65]	0.385
GIP (pg/mL)	113.27 [83.89, 151.88]	117.18 [83.91, 157.97]	111.89 [83.93, 147.52]	0.603
Glucagon (pmol/L)	6.10 [4.99, 7.67]	5.95 [4.77, 7.44]	6.27 [5.07, 7.86]	0.121
Insulin (microIU/ml)	4.20 [2.95, 5.80]	4.80 [3.60, 6.58]	3.60 [2.70, 5.50]	<0.001
Fasting Plasma glucose (mg/dl)	5.00 [4.73, 5.30]	5.00 [4.70, 5.30]	5.10 [4.80, 5.40]	0.045
Insulin resistance
HOMA IR	1.10 (0.75)	1.20 (0.63)	1.02 (0.83)	0.011
HOMA2-IR	0.55 (0.33)	0.60 (0.30)	0.50 (0.35)	<0.001

Data presented as median [interquartile range] and mean (SD) for continuous variables depending on distribution and frequencies (percentage; %) for categorical variables. P values were determined by Wilcoxon rank sum tests for continuous data and chi-square tests for categorical data. Statistically significant results are in bold.

BMI body mass index, WC waist circumference, HDL high-density lipoprotein, LDL low-density lipoprotein, GLP-1 glucagon-like peptide 1, GIP glucose-dependent insulinotropic polypeptide, HOMA-IR Homeostatic Model Assessment 2 for Insulin Resistance, HOMA2-IR Homeostatic Model Assessment for Insulin Resistance version 2.

The fasting plasma glucose levels were higher in boys (p = 0.045) and fasting insulin levels were significantly higher for girls than boys (p < 0.001), but there were no significant differences between girls and boys in median GLP-1, GIP or glucagon plasma levels.

Association analyses with z-scores of GLP-1, GIP and Glucagon

Age was positively associated with GLP-1 absolute levels in both sexes, while GIP levels showed an inverse relationship with age in girls (Fig. 1). When investigating associations with GLP-1, GIP and Glucagon z-scores in the univariate analysis in the 485 children without metabolic syndrome (Table 2) both insulin resistance surrogate markers HOMA-IR and HOMA2-IR was positively associated to GLP-1 z-score in boys (p = 0.023 and p = 0.021 respectively). There was no difference in the association with GLP-1 z-score between the sexes.

Fig. 1.

Smoothed reference percentile curves (lines) and individual observations of fasting plasma concentrations of GLP-1, GIP and

Table 2.

Associations between GLP-1, GIP and Glucagon in boys and girls.

Variable	GLP-1 z-score			GIP z-score			Glucagon z-score
	Girls	Boys	P-value	Girls	Boys	P-value	Girls	Boys	P-value
Anthropometrics
Height (cm)	−0.006 (0.158)	0.002 (0.642)	0.193	−0.000 (0.884)	0.004 (0.351)	0.671	−0.001 (0.809)	−0.000 (0.915)	0.764
	[−0.016, 0.003]	[−0.006, 0.010]		[−0.013, 0.013]	[−0.005, 0.013]		[−0.012, 0.012]	[−0.009, 0.008]
Weight (kg)	−0.008 (0.217)	0.004 (0.413)	0.125	0.004 (0.618)	0.008 (0.196)	0.684	−0.002 (0.823)	0.000 (0.976)	0.767
	[−0.019, 0.002]	[−0.006, 0.014]		[−0.011, 0.019]	[−0.004, 0.018]		[−0.018, 0.011]	[−0.015, 0.012]
BMI z-score	−0.101 (0.214)	0.023 (0.686)	0.165	0.01 (0.876)	0.088 (0.236)	0.469	−0.077 (0.374)	0.018 (0.777)	0.376
	[−0.259, 0.058]	[−0.101, 0.148]		[−0.159, 0.177]	[−0.154, 0.181]		[−0.248, 0.093]	[−0.106, 0.142]
WC z-score	−0.080 (0.339)	−0.047 (0.573)	0.168	0.096 (0.198)	0.065 (0.479)	0.729	−0.021 (0.841)	−0.027 (0.762)	0.956
	[−0.246, 0.085]	[−0.116, 0.210]		[−0.182, 0.227]	[−0.115, 0.249]		[−0.225, 0.185]	[−0.204, 0.149]
Lipids
HDL z-score	0.095 (0.232)	0.02 (0.803)	0.488	0.046 (0.613)	−0.047 (0.515)	0.413	0.057 (0.511)	0.08 (0.245)	0.853
	[−0.06, 0.25]	[−0.13, 0.17]		[−0.1897, 0.095]	[−0.188, 0.095]		[−0.11, 0.228]	[−0.053, 0.208]
LDL z-score	−0.006 (0.945)	−0.079 (0.236)	0.492	0.046 (0.561)	0.016(0.820)	0.820	0.126 (0.134)	0.063 (0.334)	0.555
	[−0.169, 0.158]	[−0.218, 0.052]		[−0.110, 0.201]	[−0.121, 0.153]		[−0.038, 0.291]	[−0.063, 0.191]
Triglycerides z-score	−0.000 (0.990)	0.062 (0.338)	0.538	0.071 (0.489)	0.027 (0.740)	0.736	0.154 (0.052)	0.115 (0.061)	0.689
	[−0.148, 0.149]	[−0.064, 0.189]		[−0.130, 0.272]	[−0.132, 0.187]		[−0.001, 0.309]	[−0.005, 0.234]
Total cholesterol z-score	0.059 (0.476)	−0.068 (0.315)	0.235	0.070 (0.357)	0.009 (0.896)	0.554	0.175 (0.032)	0.103 (0.109)	0.493
	[−0.010, 0.22]	[−0.202, 0.065]		[−0.07, 0.219]	[−0.127, 0.146]		[0.015, 0.335]	[−0.022, 0.231]
Insulin sensitivity
HOMA-IR	0.095 (0.346)	0.113 (0.023)	0.877	0.071 (0.602)	0.139 (0.168)	0.681	0.273 (0.013)	−0.027 (0.765)	0.037
	[−0.103, 0.294]	[0.016, 0.210]		[−0.195, 0.336]	[−0.06, 0.339]		[0.005, 0.491]	[−0.203, 0.148]
HOMA2	0.219 (0.322)	0.294 (0.021)	0.796	−0.179 (0.557)	0.313 (0.195)	0.731	0.613 (0.009)	0.015 (0.942)	0.067
	[−0.215, 0.654]	[0.045, 0.543]		[−0.419, 0.779]	[−0.159, 0.785]		[0.149, 1.076]	[−0.419, 0.451]

All variables were adjusted for BMI z-score, except BMI z-score, height and weight. Bold values highlight the significant results.

None of the investigated variables were associated with GIP z-scores in the univariate analysis.

For Glucagon z-scores there were a significant difference in associations between boys and girls for the surrogate marker for insulin resistance HOMA-IR (p = 0.037) and also a borderline significant difference in associations between sexes for HOMA2-IR (p = 0.067). The association with Glucagon z-score was positive for both HOMA-IR (p = 0.013) and HOMA2-IR (p = 0.090) in girls. A positive relationship was also found between Glucagon z-score and total cholesterol in girls (p = 0.032). Triglyceride z-scores were borderline significantly associated with Glucagon z-scores in both girls (p = 0.052) and boys (p = 0.061.

In the multiple linear regression analyses including sex, plasma glucose, insulin, BMI z-score, total fat mass and HOMA-IR in the model as standardized explanatory variables, we found that GIP and GLP-1 z-scores were no longer associated with any of the variables in the model. However, Glucagon z-scores where highly associated with fasting levels of insulin (p = 0.006), together with HOMA (p = 0.021) and total cholesterol (p = 0.027). Tables with age- and sex specific percentiles for GLP-1, GIP and Glucagon are to be found in Supplementary Tables S1–6.

Metabolic syndrome sub-analysis

In the cohort, 21 participants fulfilled the criteria for MetS. Those with MetS were comparable to those without MetS regarding age, sex and puberty status (Supplementary Table S7). The MetS subgroup had significantly higher BMI z-score (p < 0.001), total fat mass (p < 0.001) and higher insulin resistance by the surrogate measure HOMA2-IR (p < 0.001) than the subgroup without metabolic syndrome. The table did not include HDL, LDL, waist, blood glucose and HOMA-IR as it is included in the definition of metabolic syndrome.

No significant differences were demonstrated for GLP-1, GIP or glucagon z-scores in the group with metabolic syndrome compared to without metabolic syndrome.

Discussion

We established age- and sex- specific reference levels for the hormones GLP-1, GIP and glucagon from a Danish school cohort. To our knowledge it is the first publication providing these reference ranges as percentiles and demonstrating the large variation in the hormones during childhood and adolescence. We also demonstrate associations between incretin hormones, glucagon and metabolic parameters, highlighting both sex differences and associations with age and insulin resistance.

In recent years, we have become increasingly aware of the large heterogeneity of most metabolic diseases. It has therefore been a major interest to develop profiles for personalized medicine or treatment strategies including biomarkers for risk stratifications, which could end in a scenario where treatments are targeted to the patients that have the highest likelihood of responding. The incretin levels could potentially be a part of this scenario and that children with altered or abnormal levels might benefit from intervention. In this scenario reference intervals are essential to define these stratification cut-offs.

In our cohort girls had significantly higher values for insulin, insulin resistance (HOMA IR and HOMA2-IR), lipid profile and total fat mass compared to boys This is in accordance with previous findings where girls had higher adiposity¹⁴ and insulin resistance than boys²⁶ which was also found in this study when investigating the insulin sensitivity surrogate markers HOMA-IR and HOMA2-IR. Also, the difference in lipid profiles has been demonstrated before.⁸ In our association analysis HOMA-IR and HOMA2-IR were both associated with GLP-1 z-score in boys and glucagon z-scores in girls.

In a study investigating determinants of incretin concentrations in subjects with normal or impaired glucose tolerance; glucagon and age were both identified as being positively associated with GLP-1 concentrations in the healthy subjects²⁷

A negative association was observed between age and GIP levels in girls, but not in boys in our cohort. This finding is consistent with previous literature showing that GIP secretion is inversely related to age in adults,²³ potentially due to age-related changes in insulin sensitivity or β-cell responsiveness.²⁸ This age-related increase in absolute levels of GLP-1 and decrease in GIP levels highlights the importance of using age- and sex- specific reference percentiles to generate z-scores for GLP-1 and GIP in future research evaluating incretins levels in children.

Incretins and insulin sensitivity and puberty

In our association analysis HOMA-IR and HOMA2-IR were both associated with GLP-1 z-score in boys and glucagon z-scores in girls. The association to glucagon z-scores to HOMA-IR in girls could be due to girls being more advanced in puberty compared to boys when adjusting for age, as puberty tends to induce a state of increased insulin resistance,²⁹ higher BMI, waist circumference and body fat percentage.³⁰ However, contradicting this, in our study we found that a larger percentage of boys were being self-assessed as pubertal than girls. Though this was rather surprising it was also found in another Danish study by Rasmussen and colleagues²³ that boys tend to overestimate pubertal maturation when self-assessing compared to girls which may be an explanation to the surprising finding, but also underline that the puberty data in this study should be interpretated with caution.

However the association to insulin resistance and GLP-1 and Glucagon have been found in previous studies³¹ and when investigating the role of glucagon and GLP-1 levels in children with obesity and insulin resistance compared to a healthy group of children and adolescents; Stinson et al found that fasting levels of both GLP-1 and glucagon were elevated in the group with obesity and insulin resistance compared to the group with normal weight. This is similar to the associations found in our study. Glucagon has previously been identified as a potential biomarker signaling early glycemic dysfunction in children with adeposity.³² Stinson et al. found that elevated glucagon levels may precede the impairments seen in the glucose regulation in overweight and obese children. Elevated fasting glucagon concentrations also corresponded to worsened cardiometabolic risk outcomes in another study by the same group,³³ underscoring its potential clinical utility as a predictive biomarker. These findings emphasize the importance of establishing robust reference materials to accurately determine threshold levels at which glucagon concentrations should be considered pathologically elevated.

Incretins and obesity and metabolic syndrome

A cross-sectional study including more than 4000 participants demonstrated that fasting GLP-1 and glucagon levels were higher in children and adolescents with overweight and obesity compared to a healthy population.³¹ Likewise, GLP-1 was suggested as a potential risk indicator of metabolic syndrome in a recent study in adults.³⁴

In our study we did not find the same associations with BMI and GLP-1 or glucagon z-scores and we found no significant differences between children with metabolic syndrome compared to without metabolic syndrome in regard to GLP-1, GIP or glucagon z-scores. However, Due to the small number of participants with metabolic syndrome, our study was underpowered to detect small to moderate associations with incretins and glucagon levels. In an article by Stouwe et al. they found that fasting plasma levels of GLP-1 in young adults were positively associated with body fat mass, especially in males. This finding remained significant even after adjustment for BMI and other measures of obesity.³⁵ In our study only glucagon z-score was positively associated with total cholesterol and marginally significant for triglyceride z-score. Prospectively in a future study it would be interesting to investigate if the hormone levels could predict treatment response and could stratify into response groups before considering treatment options with GLP-1/GIP or Glucagon receptor agonists in the future. Currently, two GLP-1 RAs have been approved by the FDA for use in children and adolescents as obesity treatment from the age of 12 years. A new phase 3a trial also evaluating GLP-1 RA in children 6–12 years of age with obesity and found similar findings as for adolescents and adults including reduction in BMI and with the same adverse events.¹⁷ However, the pharmaceutical interaction with the endogenous incretin hormones during growth and development in children and adolescents has not been thoroughly investigated in any studies to our knowledge. A recent randomized controlled trial treating adolescents aged 10–18 years with obesity for 6 months with GLP-1 RA investigated the consequence on the endogenous levels of GLP-1, glucagon and insulin.¹⁸ Besides the well-known results from use of GLP-1 RA, such as significantly lowering BMI and 2-h glucose levels, this study found that the endogenous levels of GLP-1, glucagon and insulin were similar in the fasting and stimulated state for both the treatment and the placebo group. The only difference they found between the two groups were the levels of the enzyme dipeptidyl peptidase-4 (DPP-4). The role of DPP-4 is to cleave and inactivate GLP-1 within minutes from entering the circulation. As GLP-RA is synthetically designed to resist the cleaving from DPP-4, in order to increase half-life, it could explain why the DPP-4 levels were lower in the treatment group, as it might be as a result of a reduced need for cleaving. The most plausible explanation to why the endogenous levels of GLP-1 were unchanged in the study¹⁸ was that the pharmacological GLP-1 RA stimulates GLP-1 receptors in the central nervous system andreduces appetite and slows down gastric emptying, giving all the desired effects of weight loss and lowered fasting glucose – irrespective of the plasma concentrations of GLP-1 and glucagon.

Generally, very few studies have investigated incretin levels during childhood and adolescence. This is despite the large attention on GLP-1 and GIP and their therapeutic potentials to revert cardiometabolic and glycemic disturbances also in the pediatric population.³⁶

Strengths and limitations

In our study the age span represented school children and not a pediatric span from birth to adulthood (0–18 years). The cohort was well sized, but the study population has its limitations in being both geographically and demographically restricted to Danish school children, which may make it less applicable for other more heterogeneous populations in other countries. The cross-sectional design is a natural limitation making it impossible to determine causal relationships when evaluating the associations. Though it is a strength that pubertal data was included in this study—which is seldom the case in reference material–the pubertal stage was subject to bias as it was self-reported. However, as we in this study divided pubertal maturation into prepubertal and pubertal the validity becomes sufficiently accurate.²³ Adjustments for confounding factors such as pubertal stage, dietary intake or physical activity should ideally also have been taken into consideration. We only analyzed and evaluated total GLP-1 and GIP (active + metabolites) and not their active forms. It has been discussed in other papers whether to include active, total or both depending on the situation. We chose the total form, as it was commonly recommended for this kind of study and comparison.³⁷ We underline that incretin assays are only used for research purposes and that it is also essential to evaluate whether the assay used to determine incretin concentrations in other studies are comparable to ours before using the percentiles for a z-score transformation. The small sample size in the metabolic syndrome group limited the statistical power, and the risk of type II error, particularly for detecting small effect sizes.

Conclusion

In summary, our findings suggest that the regulation of incretin hormones are influenced by age, pubertal status and insulin sensitivity. Percentiles and z-scores generated by this study should be considered and used when comparing the hormones in children of different ages, pubertal and metabolic stages during their childhood development. Longitudinal studies comparing incretin and glucagon levels in healthy children compared to children with other metabolic disorders should be further investigated. Likewise, it would be of major interest to evaluate if stratified levels of GLP-1 can predict the effects of GLP-1 receptor agonists in children with metabolic disorders, insulin resistance and/or obesity in larger cohorts.

Author contributions

T.W.H.L., A.D.M., and J.J. all made substantial contributions to conception and design, acquisition of data, analysis and interpretation of data. Together with drafting the article, revising it critically for important intellectual content and made final approval of the version to be published. F.B.M., A.K.J., V.H.J., F.P., M.S.H., A.J.S., and N.W. all revised the article critically for important intellectual content and made final approval of the version to be published.

Funding

T.W.H.L. and A.D.M. have received strategic funding from Steno Diabetes Center Copenhagen. A.D.M. received funding from Herlev and Gentofte University Hospital. T.W.H.L., A.D.M., J.J., F.B.M., V.H.J., and F.P. are employed at Steno Diabetes Center Copenhagen, a public hospital and research institution under the Capital Region of Denmark, which is partly funded by a grant from the Novo Nordisk Foundation. The funders had no role in any part of this study.

Data availability

The datasets analyzed during the current study are available from the corresponding author on reasonable request.

Competing interests

The authors declare no competing interests.

Informed consent

In the original study participant and their parents (or legal guardians) gave their informed consent both verbally and written before participation. Written consent was given by the parents (or legal guardians) as the age was below 18 years.

Supplementary information

The online version contains supplementary material available at https://doi.org/10.1038/s41390-025-04551-7.