Fasting Plasma Ghrelin Levels Are Negatively Correlated With Insulin Resistance and PAI-1, but Not With Leptin, in Obese Children and Adolescents

RESEARCH DESIGN AND METHODS

For this study, 49 Japanese children and adolescents (38 boys, 11 girls) with simple obesity were recruited. Their ages ranged from 5 to 19 years (mean age 10.2 ± 2.8 years). Standing height and weight were measured, and BMI and the percent overweight were calculated. The percent overweight was determined on the basis of the Japanese standard body weights for height by age and sex using the following formula: [(actual body weight − standard weight)/standard weight] × 100%. The percent overweight values of the subjects ranged from 22.0 to 127.2% (mean value 56.0 ± 20.7%). The mean BMI was 28.0 ± 4.5. Waist circumference was also determined. The percent body fat was estimated by hand-to-foot impedance measured with a bioelectrical impedance analyzer (Sekisui Impemeter III; Sekisui, Tokyo, Japan) (13). Blood samples were collected in the outpatient clinic around 9:00 a.m. after an overnight fast, and the fasting blood glucose (FBG), immunoreactive insulin (IRI), ghrelin, leptin, and PAI-1 concentrations were measured. Plasma ghrelin concentrations were measured by a radioimmunoassay (RIA) specific for the COOH-terminal portion of ghrelin, as previously described (14). Leptin levels were measured with an RIA kit (Human Leptin RIA; Linco Research, St. Charles, MO), and PAI-1 levels were measured with an enzyme immunoassay kit (IMULYSE PAI-1; Biopool, Ventura, CA) (15). Insulin resistance was assessed by homeostasis model assessment insulin resistance index (HOMA-R) (16) and the quantitative insulin sensitivity check index (QUICKI) (17). HOMA-R was calculated as [FBG (mg/dl) × IRI (mU/l)]/405, and QUICKI, as 1/(log IRI + log FBG). This study was carried out in accordance with the principles of the Declaration of Helsinki. Informed consent was obtained from all subjects and/or their parents.

Results are given as means ± SD, unless otherwise noted. The Mann-Whitney U test and Spearman rank correlation were used for statistical analysis. P <0.05 was considered statistically significant.

RESULTS

The fasting plasma ghrelin levels were 69.1–369.8 fmol/ml. There was no significant difference between the boys (range 69.1–369.8, median 157.8 fmol/ml) and girls (range 77.8–202.2, median 137.7 fmol/ml). The circulating ghrelin levels tended to correlate negatively with age, but the correlation was not significant (r_s = −0.268, P = 0.054).

The fasting leptin levels were positively correlated with percent overweight (r_s = 0.587, P < 0.0005), BMI (r_s = 0.548, P < 0.005), percent body fat (r_s = 0.604, P < 0.0005), and waist circumference (r_s = 0.542, P < 0.0005) (Table 1, Fig. 1). The PAI-1 levels were positively correlated with the indexes of body composition, except percent body fat, but less strongly than leptin. By contrast, the fasting plasma ghrelin levels were negatively correlated with BMI (r_s = −0.317, P < 0.05) and waist circumference (r_s = −0.301, P < 0.05), but not with percent overweight or percent body fat (Table 1, Fig. 1). The fasting plasma ghrelin concentrations were not correlated with the height SD scores of the obese children and adolescents.

Table 2 shows the correlations between the indexes for insulin resistance and the fasting concentrations of ghrelin, leptin, and PAI-1. Leptin and PAI-1 were positively correlated with fasting IRI, and ghrelin was negatively correlated with fasting IRI (Table 2, Fig. 2). Fasting plasma ghrelin was negatively correlated with HOMA-R (r_s = −0.300, P < 0.05) and positively correlated with QUICKI (r_s = 0.302, P < 0.05), but the leptin levels were not correlated with either HOMA-R or QUICKI. PAI-1 was positively correlated with HOMA-R (r_s = 0.450, P < 0.005) and negatively correlated with QUICKI (r_s = −0.451, P < 0.005). It was noteworthy that there was no correlation between the fasting plasma ghrelin and fasting blood glucose concentrations (Table 2, Fig. 2).

There was also a negative correlation between the fasting plasma ghrelin and PAI-1 concentrations (r_s = −0.322, P < 0.05), but not between the ghrelin and leptin concentrations (r_s = −0.297, P = 0.40) (Fig. 3).

DISCUSSION

The plasma ghrelin levels in our study tended to correlate negatively with age, but the correlation was not significant (r_s = −0.268, P = 0.054). It may be important and interesting to clarify the effect of growth and maturation on fasting plasma ghrelin levels; however, in this study, we measured the fasting plasma ghrelin levels of only obese children and adolescents, not of nonobese subjects, and thus further study is needed to address the issue of changes in plasma ghrelin levels in association with growth and maturation.

We did not perform the glucose clamp test as the gold standard for quantifying insulin sensitivity/resistance in vivo, because it requires intravenous infusion of insulin and glucose and frequent blood samples over a 3-h period, and we could not justify applying a glucose clamp to apparently healthy children and adolescents with simple obesity. Instead we used fasting IRI, HOMA-R (16), and QUICKI (17), which are well-accepted alternatives for estimating insulin sensitivity/resistance, with QUICKI values in particular having been reported to be well correlated with indexes of insulin sensitivity in glucose clamp studies (17).

Tschöp et al. (7) has demonstrated negative correlations of fasting plasma ghrelin with percent body fat, fasting insulin concentrations, and leptin concentrations of obese adult Caucasians and Pima Indians. Furthermore, studies in normal-weight and obese humans have confirmed that serum leptin concentrations accurately reflect BMI and percent body fat (18,19). Our study showed that leptin levels were positively correlated with BMI, percent body fat, and percent overweight (Table 1, Fig. 1), but our observations of Japanese obese children and adolescents showed no correlation between fasting plasma ghrelin and percent body fat or leptin concentrations (Table 1, Figs. 1 and 3). Fasting plasma ghrelin was significantly correlated with BMI, waist circumference, fasting IRI, HOMA-R, QUICKI, and plasma PAI-1. Plasma PAI-1 levels in humans have been demonstrated to be closely correlated with visceral fat area, but not with subcutaneous fat area, and to be clearly related to insulin resistance (11,12). The plasma PAI-1 levels of our obese children and adolescents were also significantly correlated with the IRI, HOMA-R, and QUICKI values (Table 2). The characteristics of the obesity in our study population were heterogeneous in terms of insulin resistance and blood leptin and PAI-1 levels, and the differences in our results from those in the previous report on adult obese Caucasians and Pima Indians (7) may be attributable to this heterogeneity. The results of our study suggest that the downregulation of ghrelin secretion may be a consequence of a higher insulin resistance associated with visceral fat accumulation and elevated PAI-1 concentrations, and not a consequence of total body (mainly subcutaneous) fat accumulation associated with elevated leptin concentrations.

Cummings et al. (20) recently reported that plasma ghrelin levels nearly doubled immediately before each meal, and fell to trough levels within 1 h after eating, a pattern reciprocal to that of insulin levels. Those researchers suggested that single measurements of ghrelin levels during the troughs before breakfast might serve as surrogates for 24-h profiles in estimating overall ghrelin levels, as they correlated strongly with the 24-h integrated area-under-the-curve values (20). The fasting plasma ghrelin concentrations in our study were not correlated with FBG concentrations and were negatively correlated with insulin resistance. These results warrant further studies on the association between insulin resistance and higher PAI-1 concentrations and the mechanisms influencing total daily secretion of ghrelin.