Introduction

Advanced glycation end products (AGEs) form when proteins or lipids interact with reduced sugars for an extended period of time, subsequently undergoing progressive irreversible molecular transformation. Growing evidence suggested that AGEs interact with cell surface receptors for AGEs (RAGE) under hyperglycemic conditions, leading to increased oxidative stress and inflammation.1,2 Interactions of AGEs with RAGE impart distinct and maladaptive remodeling of cross-linked collagen in the vascular wall.3,4

Concentrations of serum AGEs correlate with AGEs accumulation in the skin.5 Subcutaneous AGEs seem to be related to long-term diabetic risk factors and glycemic control.5 Data on the relationship between obesity and AGEs formation are scarce and inconsistent. den Engelsen et al.6 found no association of central obesity and subcutaneous AGEs. In contrast, a more recent study demonstrated an association of skin AGEs with incidence metabolic syndrome, higher waist circumference and elevated blood pressure (BP) in adults.7 However, no association of subcutaneous AGEs with obesity was found in participants without the metabolic syndrome.7 In a recent meta-analysis of patients with high CV risk, skin AGEs has been shown to be predictive of CV and all-cause mortality.8 Sedentary lifestyle and unbalanced diet have been associated with an accumulation of AGEs.9 The combination of exercise and diet seems to be an effective means to reduce AGEs accumulation.10 A recent study has shown that life-long exercise can counteract the age-related accumulation of AGEs.11 In older adults it was recently shown that higher physical activity was associated with lower AGEs levels.12 Hypertension and arterial stiffness have both been associated with increased plasma concentrations of AGEs in adults.13

There are very few studies on AGEs in children. A previous review on the role of dietary AGEs in childhood suggested that AGEs are involved in the pathogenesis of adiposity and β-cell failure.14 Jaisson et al. found that serum AGEs were elevated on first diagnosis of diabetes mellitus and may play a role in developing long-term complications.15 Children with 5 years exposure to diabetes have been reported to have higher accumulation of skin autofluroescence (SAF)-derived AGEs, comparable to levels of healthy adults and equivalent to about 25 years of chronologic aging.16 Our study, for the first time, aimed to examine the association of BMI, body fat (BF), BP and cardiorespiratory fitness (CRF) with subcutaneous accumulation of AGEs in an unselected population of young children. We hypothesized that cardiovascular risk factors such as obesity, high blood pressure and low physical fitness were associated with higher concentration of SAF-derived AGEs early in life and that high fitness levels on the other hand would be associated with lower SAF-derived AGEs.

Methods

Study design and participants

This cross-sectional study was embedded in the large-scale, cross-sectional EXAMIN YOUTH study in Switzerland. The study protocol was approved by the Ethics Committee of the University of Basel (Switzerland, EKBB: 258/12). The study was performed in accordance with the Helsinki Declaration of Guideline For Good Clinical Practice17 and the Strengthening the Reporting of Observational Studies in Epidemiology (STROBE) guidelines.18 In a predominantly Caucasian population, children aged 6−8 years were screened for BMI, BF, BP, CRF and subcutaneous AGEs accumulation. An informed parental written informed consent was obtained from all participants. The study was registered a priori in a clinical trials registry (ClinicalTrials.gov: NCT02853747).

Measurements

Advanced glycation end products

AGEs were assessed by subcutaneous SAF. Measurements of SAF were performed using the validated AGE Reader® device (DiagnOptics Technologies BV, Groningen, Netherlands).5,19 The AGE Reader® involves an integrated spectrometer to analyze reflected excitation light. The ratio between the emission light and reflected light multiplied by 100 was used to calculate SAF, expressed in arbitrary units (au). The emission light ranges between 420 and 600 nm, whereas the reflected excitation light ranges between 300 and 420 nm. SAF allows the noninvasive, validated method to analyze AGEs in connective tissue and it is strongly correlated to AGEs accumulation in the blood serum.5 For further analysis, the arithmetic mean of three measurements at different areas at the right ventral side of the forearm was used.

Anthropometrics

Body height was measured with a wall-mounted stadiometer (Seca 206, Basel, Switzerland), weight and percentage BF were measured with the InBody device (InBody 170 Biospace device; InBody Co., Seoul, Korea). According to cutoff points for BMI (kg/m2) incorporating age and sex, children were classified in BMI categories by use of percentiles.20 Children with a BMI over the 85th percentile in their sex and age group were categorized as being overweight and over the 95th percentile as children with obesity. BP parameters were assessed using an automated oszillograph (Oscillomate, CAS Medical Systems, Branford, CT). All children were measured in a sitting position after a 5-min rest based on the recommendations of the American Heart Association.21 BP was measured five times in series and the mean of the three measurements with the smallest variation were taken for the further analysis. According to the population-based German KiGGS study, children over the 90th percentile were categorized as having high-normal BP and over the 95th percentile as children with hypertension.22

Physical fitness

The CRF assessment took place during the physical education lessons with the same equipment used for every school. After a 5-min warm-up, a 20-m shuttle run was performed. This is a well-established and validated test to measure physical fitness.23,24 A previous meta-analysis has demonstrated the feasibility and validity of the 20-m shuttle run as a surrogate marker for CRF in children.25 It is concluded that, although spiroergometry remains to be the gold standard, the shuttle run is an established alternative if a laboratory-based test is not feasible. In this progressive endurance test, the children had to run back and forth between two lines of 20 m with an initial running speed of 8.0 km/h and an increase of 0.5 km/h every minute, paced by beeps from an audio device programmed for the timing of the shuttle run test. The individual maximum was reached if the child did not cross the line for two consecutive 20-m trials within the given time, defined by the audio beeps. A 2-m range for crossing the line was allowed. The score was assessed by the numbers of stages (1 stage = 1 min) reached with a precision of 0.5 stages.

Statistical analysis

Mean SAF was analyzed across the clinical categories of BMI, BP and tertiles of CRF using univariate analysis of covariance (ANCOVA). Bivariate analysis was performed to compare clinically relevant BMI categories, physical fitness and SAF. Pearson’s correlation was used to assess the association of CRF with AGEs. Multiple linear regression models were used to compare changes in SAF with changes in BMI, BF, BP and CRF. Four different models were fitted to adjust for age and sex as well as BMI, BP and CRF. The predictor variables were preselected and, therefore, we did not perform a stepwise regression. We have selected these predefined predictors and confounders as these are classic parameters known to have an impact on the primary outcome. We found no evidence for interactions effects between the variables.

Variance homogeneity of residuals was assessed using the Tukey−Anscombe Plots. To assess normality, we used normal QQ plots of the residuals. Ninety-five percent confidence intervals were presented for measures of effect to indicate the amount of uncertainty and a two-sided level of significance of 0.05 denotes statistical significance. For analyses and graphics, an up-to-date version of Stata 15 (StataCorp LP, College Station, TX) was used. The sample size of the cross-sectional study was given by the number of children and parents giving their consent.

Results

Participants

From the 3068 children that were invited in the study, 1690 (55%) had a written consent from their parents to participate. A total of 221 children dropped out because of illness or were otherwise absent. Due to skin pigmentation or a temporary technical default of the device (AGE Reader®), 394 children had to be excluded from the data analysis, leaving 1075 children with complete measurements (Fig. 1). Age, body weight and height, BF, BMI and shuttle run data of the 615 excluded children are presented in Supplement Table S1. Excluded children had slightly higher fitness levels compared to children included in the study. Population characteristics are shown in Table 1. Based on a modified questionnaire survey,26 95% of children were Caucasian. In our cohort, 87% of children presented with normal weight (n = 934), 10% with overweight (n = 103) and 3% (n = 38) with obesity. Based on systolic BP, 77% were categorized as children with normal BP (n = 827), 9% as having high-normal BP (n = 99) and 14% as children with hypertension (n = 149). Boys were fitter (CRF: 4.0 ± 1.6 stages) but showed higher subcutaneous AGEs (SAF: 1.07 ± 0.2 au) compared to girls (CRF: 3.4 ± 1.3 stages; SAF: 1.03 ± 0.2 au, p < 0.001).

Fig. 1
figure 1

Flow diagram.

Full size image
Table 1 Population characteristics of the study.
Full size table

Group differences

The results for between group differences are shown in Table 2. Clinical BMI, systolic and diastolic BP categories were not associated with SAF in our cohort of children. Children with higher physical fitness showed lower SAF (0.99 (1.03;1.05) au) compared to children with low CRF (1.09 (1.03;1.05) au, p < 0.001). The bivariate analysis illustrates the interrelation between fitness, BMI and SAF levels. Children with low fitness and overweight or obesity had the highest subcutaneous AGEs levels (Fig. 2).

Table 2 Advanced glycation end products in relation to clinical categories of body mass index, blood pressure and tertiles of shuttle run.
Full size table
Fig. 2
figure 2

Skin autofluorescence in relation to body mass index categories and median of physical fitness.

Full size image

Regression analysis

In the regression analysis, BMI was associated with SAF (0.006 (0.2E-3;0.011) au, p = 0.042, R2 = 0.004) with little evidence for an association after adjustment for age and sex (Table 3). Percentage BF was associated with increased SAF (0.003 (0.001;0.004) au, p = 0.001, R2 = 0.026). After adjustment for BP and CRF, no statistically significant association was found. Likewise, there was little evidence for an association of systolic and diastolic BP with SAF. One stage increase in shuttle run was significantly associated with decreased SAF, independent of BMI and BP (Fig. 3a, b). CRF alone explained 5% of the SAF variance. As expected, children with higher CRF had lower BMI (p < 0.001) and lower systolic (p < 0.001) and diastolic BP (p = 0.035).

Table 3 Regression analysis for the association of body composition, peripheral blood pressure, physical fitness and activity with advanced glycation end products.
Full size table
Fig. 3: Skin autofluorescence and cardiorespiratory fitness.
figure 3

a, b The association of cardiorespiratory fitness with skin autofluorescence. CI confidence interval.

Full size image

Discussion

This is the first study to examine the association of BMI, BF, BP, and CRF with SAF-derived AGEs accumulation in children. Our findings demonstrate that low physical fitness is associated with increased accumulation of subcutaneous AGEs in young children. In contrast to our hypothesis, we found no independent association of BMI, BF and BP with SAF-derived AGEs.

Physical fitness and AGEs

Higher CRF was associated with reduced subcutaneous AGEs formation, indicating a favorable glucose metabolism and a reduction of associated CV risk in these children. The confidence intervals show evidence for and association of CRF with SAF-derived accumulation of AGEs in young children. Data in a Slovak population suggested that regular self-reported physical activity is associated with lower SAF during lifespan.27 Two studies have previously measured subcutaneous AGEs in a small number of healthy young children.27,28 Our study is the first to assess SAF-derived AGEs in a large population-based unselected cohort of 6−8-year-old children, offering reliable normal values for young Caucasian children (mean SAF 1.05 ± 0.20 au) and demonstrating the inverse association with objectively measured CRF.

CRF affects glycation processes in children, as at least 5% of the variance of SAF was explained by CRF in our cohort. One unit increase in CRF was associated with a 0.03 au decrease in SAF and the difference in SAF between the lowest and the fittest tertile was 0.10 au. In comparison, in adolescents with type 1 diabetes one unit increase in hemoglobin A1c (HbA1c) has been associated with a 0.06 au increase in SAF.29 In their cohort, the difference in SAF between type 1 diabetes and healthy controls was 0.26 au. In light of this evidence, it may be concluded that our findings are potentially of clinical relevance. Future studies will have to determine the long-term clinical and predictive value of subcutaneous AGEs for the development of cardiometabolic disease and the potential of physical fitness to counteract accumulation of AGEs during childhood and later in life.

Body mass index, body fat and AGEs

No independent association between BMI and SAF was found in our cohort of children. The association of BF with SAF was not independent of BP and CRF.

A previous study in adults also showed no association of subcutaneous AGEs with obesity in the absence of the metabolic syndrome.7 A prior study in children suggested that BMI, BF and fat mass were associated with soluble RAGE in older children aged 12−14 years.30 It is therefore possible that obesity only affects SAF-derived AGEs formation after a longer-term exposure time to an increased BMI. In addition, sex seems to be a nonmodifiable factor for the SAF-derived accumulation of AGEs in children. In our population of young children, boys showed higher subcutaneous AGE accumulation compared to girls independent of BMI and BP. In contrast to the gender differences in our children, it has been shown that plasma accumulation of AGEs is higher in women compared to men aged around 20 years.31 Our children were examined in pre-puberty, whereas the aforementioned study investigated young adults. Sex-related differences in childhood development and puberty seems to be the most likely explanation for this conundrum.

Blood pressure and AGEs

In the regression analysis, no association of systolic and diastolic BP with SAF was found. However, there was a weak but significant association of diastolic BP and SAF in children categorized as children with high-normal BP and hypertension. In adults, hypertension has been associated with increased accumulation of AGEs in plasma.13 A recent study found an association of systolic and diastolic BP with subcutaneous AGEs in a general adult population.32 In patients with the metabolic syndrome, high-normal BP was also associated with subcutaneous AGEs.7 Childhood BP has been shown to predict development of CV disease in adulthood.33 In children, higher BP does not seem to directly and independently affect SAF-derived AGEs accumulation and metabolic health. As argued before, exposure time to high BP may not be long enough to affect SAF-derived AGEs accumulation in young children. Based on our findings the clinical application of subcutaneous AGE’s in young children to differentiate cardiometabolic risk would appear premature. It remains to be determined if and to what extent BP, and indeed BMI, affect SAF-derived AGEs accumulation in older children and adolescents.

Potential mechanisms

Endothelial dysfunction and obesity-related inflammation are mediated through oxidative stress conditions.34 It is well known that oxidative stress is a main determinant for increased formation of AGEs.35,36 Sedentary behavior is characterized by reduced mitochondrial capacity and increased oxidative stress and exercise has the potential to reverse oxidative conditions.37 Proteins are glycated to form AGEs through the so-called Maillard reaction. Early non-enzymatic glycation and formation of Schiff bases and Amadori products represent reversible cross-links between proteins and sugars.38 We hypothesize that exercise can reverse the formation of early glycation products preventing irreversible cross-links and tissue accumulation of AGEs forming fluorescent derivatives. In addition, exercise-induced formation of soluble RAGE may play an important role in reducing AGEs accumulation and associated oxidative stress. The circulating soluble RAGE binds to AGE and acts as a competitive inhibitor of ligands that activate RAGE. Long-term physical activity and exercise lead to an increase in soluble RAGE, which blocks RAGE activation.39

Improvement of AGEs metabolism may be achieved by physical fitness interventions rather than measures focusing on classical risk factors such as BMI and BP reduction. From a pathophysiological point of view, it is possible that other sensitive metabolites of the AGEs metabolism, such as RAGE or protein-bound AGEs and markers of dicarbonyl stress, may be associated with BMI and BP in young children. SAF-derived AGEs accumulation may occur at later stages compared to increases in circulating serum and urine biomarkers of AGEs metabolism. Future studies will have to analyze blood or urine samples of young children to clarify the clinical relevance and differences of circulating AGEs metabolites as compared to subcutaneous AGEs accumulation.

Strengths and limitations

This is a cross-sectional design and does not investigate temporal development of the associations. However, an inverse relationship between CRF and SAF-derived AGEs was found and a long-term follow-up is warranted to proof causal associations between lifestyle-related risk factors and metabolic health in children. Furthermore, only 3% were children with obesity in our cohort. Studies in populations with a higher prevalence of children with obesity may help to further differentiate the association of obesity and AGEs in children. CRF was assessed by a 20-m shuttle run and not by spiroergometry as the gold standard for objective assessment of physical fitness. Twenty-meter shuttle run, however, is an appropriate surrogate if spiroergometry is not available. It is a valid tool to assess CRF in all children including children with obesity.25,40 With respect to measuring SAF-derived AGEs, the device cannot be applied in children with dark skin and, therefore, a selection bias is given for technical reasons. Of note, the measurement of SAF-derived AGEs only detects complex mixtures of AGE and non-AGE fluorophores in a qualitative manner. This cross-sectional study is limited to explaining PA-mediated mechanisms of action on SAF-derived AGEs. Interventional studies are warranted, which ideally include the full spectrum of AGEs metabolites in serum. This was beyond the scope of our population-based approach. The R2 values show that there is a high variability in our cohort and a large sample size is needed to confirm our findings. However, the CIs indicate that there is evidence that CRF is associated with SAF-derived accumulation of AGEs. This result was independent of BMI and blood pressure. Only about 7% of the variation in SAF can be explained by a single factor such as CRF. The mechanisms involved are likely to be multifactorial as is often the case in complex physiological systems. Future longer-term studies have to investigate if this association is predictive for the development of cardiometabolic disease later in life. One strength of our study includes the large sample size and the use of standardized procedures to measure BMI, BF, BP and CRF in children.

Conclusions

In conclusion, our results demonstrate that physical fitness was associated with SAF-derived AGEs in young children without evidence for statistically significant associations with body mass and BP. Analysis of skin AGEs is a feasible tool to differentiate the effects of physical fitness on tissue glycation and metabolism. We postulate that exposure times to a higher BMI and BP are too short to affect AGEs deposition in tissue of young otherwise healthy children. Associations may still become apparent at later stages and during adolescents. From a clinical perspective, higher AGEs have been linked with development of diabetes mellitus in children15,16 but, as our findings demonstrate, this is not the case for childhood obesity and hypertension. In children with increased AGEs and referenced to our normal values, treatment strategies should still focus on reducing AGEs accumulation to counteract the growing prevalence of metabolic disease in childhood and later in life. Primary prevention programs may need to focus on improving physical activity and fitness as treatment options to achieve this ambitious long-term goal.