Ghrelin, Adipokines, Metabolic Factors in Relation With Weight Status in School-children and Results of a 1-year Lifestyle Intervention Program

Christine Rambhojan; Elodie Bouaziz-Amar; Laurent Larifla; Jacqueline Deloumeaux; Josiane Clepier; Jean Plumasseau; Jean-Marc Lacorte; Lydia Foucan


Nutr Metab. 2015;12(43) 

In This Article


Analysis of Parameters at Baseline

Overall 120 children of both sexes were enrolled in the study.

The accompanying parents were 80 % women and 20 % men. Forty four (37 %) of them were overweight/obese (BMI > 25 kg/m2) and 29 % were obese (BMI ≥ 30 kg/m2). Family history of diabetes was noted in 50 % of accompanying parents and 7 % had personal history of diabetes.

The children were categorized into three groups: group 1 (G1: normal weight; n = 44), group 2 (G2: overweight; n = 39) and group 3 (G3: obeses; n = 37). The characteristics of the study population according to weight status at baseline are presented in Table 1.

The mean age of the study population was 12.4 ± 1.1 years and 59 % were girls. Tanner stages were distributed as follows: stage 2: 7.5 %, stage 3: 58.3 %, stage 4: 30 % and stage 5: 4.2 %.

Overall, 92.5 % of the children had a tanner stage ≥ 3 (pubertal/post pubertal development) with no significant difference between the three groups (P = 0.82, Table 1). The frequency of children eating at school canteen was higher in normal weight (53 %) than in overweight (28 %) and obese (30 %) children; (P = 0.027), (data not shown).

In accordance with the design of the study, all the anthropometric parameters (BMI, BMI z-score, WC and WC/height) were significantly different between the three groups (P < 0.001 for all parameters). Obese children had higher SBP (P = 0.015), triglycerides (P = 0.007), fasting blood glucose (P = 0.019), insulin (P < 0.001) and leptin levels (P < 0.001), whereas they had lower HDL-C (P < 0.001), ghrelin (P = 0.001) and adiponectin levels (P < 0.001) than the other groups. Abdominal obesity and insulin resistance were also more frequently found in the obese group (P < 0.001 for both).

Leptin levels were higher in girls than in boys in the overall study sample (27 vs 16 ng/mL; P = 0.004) whereas no significant gender differences were found for adiponectin (P = 0.164) and ghrelin levels (P =0.110).

The correlations between leptin, ghrelin, adiponectin levels and the other continuous variables: anthropometric parameters, blood lipids and insulin levels, among boys and among girls respectively are presented in Table 2a and Table 2b. Leptin, ghrelin, adiponectin levels were significantly correlated with the three anthropometric parameters (BMI z score, WC, WHt) in both genders except for ghrelin in girls. Concerning FBG and blood lipid parameters, in boys, significant correlations were found except for ghrelin with TG levels, for leptin with FBG and LDL-C and for adiponectin with FBG, LDL-C and TG levels. In girls, no significant correlation was found for ghrelin with the 4 parameters, for leptin with LDL-C and for adiponectin with LDL-C and TG.

High significant correlations were also noted between ghrelin, leptin and adiponectin levels with insulin levels and HOMA-IR in both genders except for ghrelin levels in girls.

All anthropometric parameters were positively correlated with insulin and HOMA-IR in both genders (data not shown).

In considering the weight status of accompanying parents, insulin resistance (HOMA IR > 3.16) was noted in 61 % of children from parents with normal weight and in 70 % of children from overweight or obese parents (data not shown).

Table 3 presents the results of the multivariate linear regression analyses exploring the associations of ghrelin, adiponectin and leptin with age, gender, weight status, HOMA-IR and tanner stage as independent variables, at baseline.

In model 1, for ghrelin levels, no significant association was noted with the five variables but association with HOMA-IR was nearly significant (P = 0.070). The model accounted for 15 % (r 2 = 0.15) of the variability in ghrelin levels.

In model 2, for adiponectin levels, significant negative associations were noted with overweight (P = 0.002), obesity (P = 0.001) and HOMA-IR (P = 0.010) and the model accounted for 26 % (r 2 = 0.26) of the variability in adiponectin levels.

In model 3, for leptin levels, significant positive associations were noted with gender (P = 0.011), overweight (P < 0.001), obesity (P < 0.001) and HOMA-IR (P = 0.002). The average increase in leptin levels was 7.9 ng/dL for gender (girls), 13.9 ng/dL for overweight, 26.4 ng/dL for obesity and 1.50 ng/dL for each unit of HOMA-IR. This multivariate model accounted for 58 % (r 2 = 0.58) of the variability in leptin levels.

No significant average change in ghrelin, leptin and adiponectin levels was noted in relation with tanner stage.

Table 4 presents changes in clinical and metabolic parameters at one year of follow-up. Of the 120 children who participated to the study at baseline, 83 have accepted to undergo evaluation for clinical and biological parameters after one year (28 in normal weight, 28 in overweight, and 27 in obese children).

Mean BMI z-score significantly decreased in overweight (1.47 to 1.33; P = 0.021) and obese children (2.18 to 2.08; P = 0.008) whereas, no significant change was noted in the normal weight group (−0.10 to −0.16; P = 0.446) (Table 4).

Concerning the metabolic parameters, significant decreases were observed in normal weight children for LDL-C (P = 0.018) and insulin levels (P =0.048). The HDL-C levels decreased in this group (P <0.001) but remained higher than in the other groups. Fasting blood glucose (P = 0.047), LDL-C (P = 0.015) and insulin levels (P = 0.019) improved in overweight children. In obese children, significant improvements were also observed in FBG (P = 0.002), insulin levels (P = 0.019) and HOMA-IR (P = 0.011). Significant rises in adiponectin levels were observed in normal weight overweight and obese children (P = 0.001, P = 0.002, P < 0.001 respectively). Leptin levels remained unchanged at 1 year of follow-up in normal weight and overweight children but decreased significantly in obese children (P = 0.008). Table 4 and fig. 1 present changes in clinical and metabolic parameters at one year of follow-up.

Figure 1.

Changes in HOMA-IR index, leptin and adiponectin levels in normal-weight, overweight and obese children after a 1 year of lifestyle intervention. HOMA-IR normal weight: P = 0.065, overweight: P = 0.079, obese: P = 0.011. Leptin levels normal weight: P = 0.234, overweight: P = 0.857, obese: P = 0.008. Adiponectin levels normal weight: P = 0.001, overweight: P = 0.002, obese: P < 0.001

Combining the results of the three groups, a nearly significant correlation was found between changes in BMI z-score, and changes in leptin concentrations (r = 0.20; P = 0.060). This positive correlation was significant in the obese group (r = 0.39; P = 0.049). Conversely no significant correlation was found between changes in BMI z-score, and both changes in HOMA-IR and adiponectin levels (data not shown).

We evaluated the frequencies of children undergoing concomitant changes in the three metabolic parameters at 1 year (decrease HOMA IR and leptin levels and increase adiponectin levels) and we found a frequency of 27 % in children from parents with normal weight and of 37 % in children from overweight or obese parents.