Circulating Levels of FGF-21 in Obese Youth: Associations With Liver Fat Content and Markers of Liver Damage

Abstract

Objective:

Fibroblast growth factor (FGF)-21 is highly expressed in the liver and regulates glucose and lipid metabolism in rodents. The effects of obesity and fatty liver on circulating FGF-21 levels have been described mainly in adults. Herein, we measured plasma FGF-21 levels in lean and obese adolescents with low and high hepatic fat content (HFF% <5.5% and HFF% ≥5.5%, respectively) and explored their relationship with hepatic fat content, measures of hepatic apoptosis, and insulin sensitivity.

Methods:

A total of 217 lean and obese adolescents with both low and high HFF% (lean = 31; obese low HFF% = 107; and obese high HFF% = 79) underwent an oral glucose tolerance test, a fast gradient magnetic resonance imaging to measure the %HFF and abdominal fat distribution. Cytokeratin 18 levels were measured as a biomarker of liver apoptosis. A subset of adolescents underwent a 2-step hyperinsulinemic-euglycemic clamp, and a liver biopsy (N = 14), to assess insulin sensitivity and steatohepatitis, respectively.

Results:

Compared to controls, FGF-21 levels were higher in obese youth, especially in those with high HFF (P < .001). FGF-21 significantly correlated with adiposity indexes (P < .001), visceral fat (r² = 0.240, P < .001), hepatic fat content (r² = 0.278, P < .001), cytokeratin 18 (r² = 0.217, P < .001), and alanine aminotransferase (r² = .164, P < .001). In subjects with steatoheaptitis, FGF-21 levels significantly correlated with the nonalcoholic fatty liver disease activity score (r² = 0.27, P = .04). Stepwise regression analysis indicated that these relationships are independent of body mass index, visceral fat, and insulin sensitivity. An inverse correlation was documented with insulin, hepatic resistance indexes, and adipose resistance indexes, which disappeared after adjusting for hepatic fat content.

Conclusions:

Plasma FGF-21 levels are increased in obese adolescents, particularly in those with fatty liver. FGF-21 concentrations significantly and independently correlate with hepatic fat content and markers of hepatic apoptosis in obese youths.

Fibroblast growth factors (FGFs) play key roles in the regulation of many cellular processes ranging from development to survival in mammals (1, 2). To date, 22 FGF members have been discovered. Among them, FGF-19 (FGF-15 in mice), FGF-21, and FGF-23 were found to similarly lack a conventional heparin binding domain, which allows them to reach the circulation and to act as hormones (1, 2). Instead of heparin, FGF-19, FGF-21, and FGF-23 utilize Klotho cofactor proteins for binding to and activation of fibroblast growth factor receptors. FGF “endocrine” subgroup, and especially FGF-21, have emerged as important endocrine factors involved in glucose and lipid metabolism and energy regulation (3, 4). FGF-21 is primarily expressed by liver, adipose tissue, and pancreas, whereas the enterocytes in the distal part of the small intestine are the main cells producing FGF-19, following the consumption of a meal (5). Both FGFs mediate their effects on tissues through different classic FGF cell surface receptor isomers (1–4, 6, 7). FGF-21 stimulates glucose uptake in adipocytes (7) and regulates energy metabolism and enhanced mitochondrial oxidative function through the activation of AMP-activated protein kinase and sirtuin 1 (8). Transgenic mice that overexpress FGF-21 had higher glucose clearance and were resistant to diet-induced obesity (7). Administration of recombinant FGF-21 to diabetic fatty rats (7) and diabetic rhesus monkeys (9) reduced plasma glucose and triglycerides, lowered low-density lipoprotein cholesterol, increased high-density lipoprotein cholesterol, and caused a modest weight loss. More importantly, FGF-21 mRNA expression in the human liver increases with steatosis grade, and its serum level is significantly increased in adult nonalcoholic fatty liver disease (NAFLD) patients (10).

Associations of this circulating hormone with glucose, insulin, and lipids metabolism as well as hepatic fat content have been described to a large extent in animals, whereas the physiological roles of FGF-21 in humans remain insufficiently understood, particularly in childhood. Although FGF-21 levels have been reported to be paradoxically elevated in adult subjects with fatty liver, one study in obese children showed no difference with young subjects with or without ultrasound-defined NAFLD (11). Therefore, to gain insights in the putative emerging role of FGF-21 in the pathogenesis of NAFLD, we measured circulating plasma FGF-21 levels in a relatively large group of obese youth with low and high hepatic fat content (HFF% <5.5% and HFF% ≥5.5%, respectively) compared to healthy nonobese peers. Macrovesicular liver fat content was measured noninvasively by the fast-magnetic resonance imaging (MRI) in the entire cohort and in a subset of obese youth by a liver biopsy. We further examined the relationships between FGF-21 with both direct and indirect measures of insulin sensitivity, hepatic insulin sensitivity, and suppression of lipolysis in adolescents with varying degrees of obesity and liver fat content.

Materials and Methods

A group of 217 obese and lean adolescents (N = 186 and N = 31, respectively) was recruited from the Yale Pediatric Obesity Clinic and are part of a longitudinal study on the pathophysiology of type 2 diabetes. Obesity and lean subjects were defined by body mass index (BMI) >95th and by BMI <87th percentile for age and gender, respectively (http://www.cdc.gov/GrowthCharts). Subjects with medical conditions or using medications at the time of recruitment that may affect lipid or glucose metabolism were excluded. All subjects were nonsmokers. Information defining the absence of alcohol consumption was obtained using a questionnaire. Autoimmune hepatitis, Wilson disease, α-1-antitrypsin deficiency, hepatitis B and C, and iron overload were excluded in subjects with persistent elevation in alanine aminotransferase.

All subjects underwent an oral glucose tolerance test as previously described (12). Indexes of insulin sensitivity (whole body insulin sensitivity index, WBISI) (12) and glucose tolerance (normal glucose tolerance; impaired glucose tolerance; type 2 diabetes) according to American Diabetes Association criteria were defined (13). Fasting samples were also collected for measuring FGF-21 and cytokeratin 18 (CK18) levels as a marker of liver apoptosis (14). In all subjects total body composition was measured by dual-energy x-ray absorptiometry and the fast gradient magnetic resonance imaging was used to measure the HFF% and abdominal fat mass distribution. Based on the analysis of hepatic fat content (Table 1), all subjects were divided into the following 3 groups: 1) lean controls (N = 31) with low hepatic fat content (HFF < 5.5%); 2) obese adolescents (N = 107) with low hepatic fat content (HFF < 5.5%); 3) obese adolescents (N = 79) with high hepatic content (HFF ≥ 5.5%).

Table 1.

Main Clinical Characteristic and Body Fat Distribution in Lean and Obese Youth with Low and High Hepatic Fat Content

	Lean	Obese with Low Hepatic Fat Content	Obese with High Hepatic Fat Content	P for Trenda
Numerosity	31	107	79
GT (NGT/IGT/T2D)	31/0/0	83/22/2	41/33/5	.001
Ethnicity (C/AA/H)	17/7/7	47/40/20	33/11/35	.001
Gender, M/F	14/17	38/69	39/40	.154
Anthropometric
Age, y	15.7 ± 0.5	14.7 ± 0.3	14.5 ± 0.4	.192
BMI, kg/m2	21.1 ± 0.3	34.04 ± 0.6	35.7 ± 0.6	<.001b,c,d
BMI-Z	0.32 ± 0.11	2.12 ± 0.05	2.3 ± 0.03	<.001b,c,d
FM, %	21 ± 1.0	41.8 ± 0.8	45.5 ± 0.8	<.001b,c,d
LBM, kg	45.0 ± 1.8	52.3 ± 1.1	52.7 ± 1.7	.016b,c
FM, kg	11.8 ± 0.7	39.1 ± 1.5	44.6 ± 1.6	<.001b,c,d
Body fat distribution
Abdominal
Visceral fat, cm2	20 ± 2	52 ± 2	74 ± 3	<.001b,c,d
Subcutaneous fat, cm2	136 ± 10	489 ± 21	560 ± 21	<.001b,c,d
Visc/sub ratio	0.129 ± 0.008	0.105 ± 0.005	0.129 ± 0.006	<.001b,d
Deep subQ, cm2	43.7 ± 4.5	163 ± 8	195 ± 8	<.001b,c,d
Superficial SubQ, cm2	49 ± 3	140 ± 5	137 ± 7	<.001b,c
Liver fat content, %	0.48 ± 0.19	1.31 ± 0.16	18.90 ± 1.37	<.001c,d
Glucose metabolism
Fasting insulin, mcU/mL	14.90 ± 1.19	28.56 ± 1.98	43.63 ± 2.94	<.001b,c,d
Fasting C-peptide, pmol/L	660.69 ± 33.86	1030.26 ± 50.33	1360.42 ± 65.81	<.001b,c,d
Liver enzymes
ALT, UI/L	14 ± 1	28 ± 5	48 ± 4	<.001c,d
AST, UI/L	20 ± 1	26 ± 2	34 ± 2	.006c,d
Marker of apoptosis
CK18, UI/L	115.9 ± 4.4	133.5 ± 5.6	200.9 ± 18.1	.004c,d

Abbreviations: AA, African American; AST, aspartate aminotransferase; BF, body fat; C, Caucasian; F, female; FM, fat mass; GT, glucose tolerance; H, Hispanic; IGT, impaired glucose tolerance; LBM, lean body mass; M, male; NGT, normal glucose tolerance; T2D, type 2 diabetes; Values are expressed as mean ± SEM.

^aP adjusted: age, gender, ethnicity.

^bP < .05 lean vs obese with low hepatic fat content.

^cP < .05 lean vs obese with high hepatic fat content.

^dP < .05 obese with low hepatic fat content vs obese with high hepatic fat content.

A subset of subjects (118 adolescents: 7 lean, 65 obese HFF <5.5%, and 46 obese HFF ≥5.5%) underwent a 2-step hyperinsulinemic-euglycemic clamp with coinfusion of 6,2d-glucose a 5d-gycerol to further assess muscle, hepatic, and adipose insulin sensitivity, respectively, as previously reported (15). In 14 subjects, due to persistent elevations in alanine aminotransferase (ALT) levels, a liver biopsy was obtained that revealed steatohepatitis (NASH). The study was approved by the Yale University Human Investigation Committee. Written informed consent was obtained from the parents and assent from the children and adolescents.

Biochemical analysis

Samples were stored at −80°C until the time of the analysis of serum FGF-21, which was measured using ELISA (quantitative human FGF-21 ELISA; R&D Systems, Minneapolis, Minnesota). The intra-assay and interassay coefficients of variability were 2.8 and 7.0%, respectively. The assay has a minimum limit of detection of 4.6 pg/mL. No samples were below the lower limit of detection.

Plasma glucose was measured using the YSI 2700 STAT Analyzer (Yellow Springs Instruments, Yellow Springs, Ohio) and lipids were measured using an Autoanalyzer (model 747-200; Roche-Hitachi). Plasma insulin and C-peptide levels were measured using double antibody RIA from Millipore (Billerica, Massachusetts) and from Diagnostic Products Corp. (Los Angeles, California), respectively. The intra-assay variation was 4.5% for insulin and 5.9% for C-peptide, and the interassay variation was 10% for insulin and 11% for C-peptide. Liver enzymes were measured using standard automated kinetic enzymatic assays. CK18 was measured using immune-based (ELISA) assay as previously reported by Feldstein et al (14, 16, 17). Plasma-free fatty acids were assayed by a colorimetric method.

Imaging studies

Abdominal MRI and total body composition (dual-energy x-ray absorptiometry)

Multislice abdominal MRI studies were performed on a Siemens Sonata 1.5 T system (Siemens Medical Solutions, Malvern, Pennsylvania) (18). Total body composition was measured by dual-energy x-ray absorptiometry with a Hologic scanner (Boston, Massachusetts).

Fast-MRI: Liver fat content

Measurement of liver fat content was performed by MRI using the 2-point Dixon method as modified by Fishbein et al (19) and reported by Cali et al (20). The fast-MRI was validated against ¹H-nuclear magnetic resonance as previously reported (21).

Liver histology

The histological diagnosis of NAFLD was established by study pathologists according to their expertise. The NAFLD activity score was determined for each patient as previously described (22). In detail, steatosis was assessed as the percentage of hepatocytes involved within a lobule (0%–100%, steatosis score) and by using a 4-grade classification modified from Kleiner et al (22): 0, absent; 1, <5%; 2, 5%–33%; 3, 33%–66%; 4, >66%. Staging and grading were performed according to Brunt et al (23). The NAFLD activity score (NAS) was calculated according to Kleiner et al (22). The score is defined as the sum of the scores for steatosis (0, <5%; 1, 5%–33%; 2, 33%–66%; 3, >66%), lobular inflammation (0, none; 1, <2 foci/×200 magnification field; 2, 2–4 foci/×200 magnification field; 3, >4 foci/×200 magnification field), and ballooning (0, none; 1, few; 2, many). Thus, this score ranges from 0 to 8. According to this classification, NASH is defined by a NAS of 5 or more.

Statistical analysis

The distribution of continuous variables was examined for skewness and kurtosis and parameters were logarithmically transformed, when appropriate. Data are expressed as means ± SD or SE. Differences in gender were analyzed by Fisher exact test, whereas dichotomous variables were analyzed by χ² test.

Differences across the 3 groups (lean, obese with low hepatic fat at baseline, and obese with high hepatic fat at baseline) were analyzed by one-way ANOVA test and post-hoc pairwise comparisons were made, adjusting the level of significance for multiple comparisons (age, gender, ethnicity, and body mass index Z score [BMI-Z]) with the Bonferroni correction. To evaluate the relationship between FGF-21 concentrations and liver steatosis as well the main metabolic and indexes of hepatic apoptosis, a Pearson correlation was performed, adjusting for age, gender, ethnicity, and BMI-Z. In the subgroup of subjects who underwent a liver biopsy, a Pearson correlation was also performed to explore the relationship between FGF-21 levels and NAS score. To further characterize the relationship between FGF-21 concentrations and obesity, a linear regression was performed using FGF-21 as dependent variables, whereas BMI, visceral fat content, age, gender, and ethnicity were used as independent variables. Finally, to characterize the relationship between FGF-21 concentrations and both liver steatosis and apoptosis further, a linear regression was also performed using 2 models using HFF and CK18 values, respectively, as the dependent variables. In the models, FGF-21, BMI-Z, WBISI, visceral fat content, age, gender, and ethnicity were used as independent variables. All analyses were performed using SPSS 19.0 for Windows (SPSS, Inc, Chicago, Illinois) and statistical significance was assessed at the 2-tailed 0.05 threshold.

Results

Anthropometric and circulating FGF-21 levels: Effects of obesity and hepatic fat content (Table 1 and Figure 1)

Figure 1.

FGF-21 concentrations in lean and obese youth with low and high hepatic fat content.

As shown in Table 1, based on BMI, the cohort was stratified into a lean and obese group; the latter was than further divided into 2 groups of obese with normal and high hepatic content. Age and gender distributions across the 3 groups were similar. As expected, a significant and progressive increase in BMI, BMI-Z, body fat, lean body mass, and fat mass was found across lean, obese with normal, and obese with high hepatic fat content. The visceral, subcutaneous, deep, and superficial subcutaneous significantly and progressively increased across the 3 groups. By design, hepatic fat content was significantly increased across the 3 groups (Table 1). These differences in the hepatic fat content were paralleled by higher levels of liver enzymes and CK18 levels across the 3 groups (Table 1). Fasting and 2-hour glucose levels increased significantly across the 3 groups of adolescents, as did fasting insulin, and C-peptide (data not shown). However, no difference was found in terms of fasting glucose between lean and obese adolescent with low hepatic fat content. Insulin stimulated glucose disposal (“M” = mg/lean body mass/min) was significantly lower across the 3 groups (see Supplemental Table 1, published on The Endocrine Society's Journals Online web site at http://jcem.endojournals.org). Fasting hepatic glucose production was significantly greater in those adolescents with fatty liver as well as percentage of FFA suppression and the hepatic and adipocyte resistance index (see Supplemental Table 1).

With the increase in both obesity and hepatic fat content, FGF-21 levels increased progressively in lean (99 ± 12 pg/mL), obese with low (135 ± 8 pg/mL), and obese with high hepatic fat content (277 ± 21 pg/mL, P < .001) (Figure 1).

Relationship between FGF-21 and obesity (Table 2)

Table 2.

Stepwise Linear Regression Analysis for Predicting FGF-21 Levels in the Entire Cohort

Independent Variables	FGF-21
	β	P	β	P
Age, y	−.012	.866	−.067	.290
Gender	−.004	.950	−.031	.630
Ethnicity	.005	.946	.014	.831
BMI-Z	.351	<.001
Visceral fat content, cm2			.497	<.001
R2	.123		.247

To characterize the relationship between FGF-21 concentrations and obesity further, a linear regression was performed using FGF-21 as dependent variables, whereas BMI, visceral fat content, age, gender, and ethnicity were used as independent variables. As shown in Table 2, BMI degree and visceral obesity are strongly related to FGF-21 levels.

Relationships between FGF-21, hepatic fat content, and biomarkers of liver apoptosis and injury (Figure 2 and Table 3)

Figure 2.

Correlation between FGF-21 and visceral fat, HFF%, CK18, and ALT in the study cohort.

Table 3.

Stepwise Linear Regression Analysis for Predicting Hepatic Fat Content in the Entire Cohort

Independent Variables	β	P	β	P	β	P	β	P	β	P
Hepatic fat content
Age, y	−.016	.819	.034	.606	−.131	.056	−.091	.150	−.105	.092
Gender	−.027	.681	−.091	.156	−.034	.567	−.052	.392	−.085	.143
Ethnicity	.067	.317	.028	.659	.079	.184	.096	.119	.076	.197
BMI-Z	.409	<.001	.176	.028	−.070	.450	.236	<.001	−.100	.254
WBISI			−.381	<.001					−.200	.004
Visceral fat content, cm2					.562	<.001			.294	<.001
FGF-21, pg/mL							.455	<.001	.312	<.001
R2	.167		.259		.289		.340		.431
CK18
Age, y	.004	.965	.058	.522	−.018	.841	−.052	.533	.005	.959
Gender	−.095	.270	−.098	.296	−.088	.310	−.077	.362	−.108	.204
Ethnicity	.161	.065	.126	.152	.149	.085	.177	.034	.079	.374
BMI-Z	.270	.002	.119	.262	.064	.593	.110	.208	.001	.998
WBISI			−.305	<.001					−.201	.023
Visceral fat content, cm2					.303	.001			.010	.928
FGF-21, pg/mL							.411	<.001	.373	<.001
R2	.073		.093		.092		.210		.215

FGF-21 levels were directly correlated with hepatic fat content (r² = 0.278, P < .001), marker of hepatic apoptosis CK18 (r² = 0.217, P < .001), and ALT (r² = 0.164, P < .001).

To investigate the association of FGF-21 further with more advanced stages of fatty liver disease such as steato-hepatitis, we analyzed the data obtained in a subgroup of subjects who underwent a liver biopsy. Circulating FGF-21 levels were significantly and positively associated with NAS score (r² = 0.27, P = .04).

To investigate the relationship between FGF-21 and hepatic fat content and markers of hepatic apoptosis further, a multiple stepwise analysis was performed. As shown in Table 3, a significant association between the FGF-21 and both hepatic fat content, defined by HFF values, and hepatic cells apoptosis, as defined by CK18 concentrations, was found in the entire cohort, even after adjusting for relevant confounding factors, confirming a relevant contribution of the FGF-21 to hepatic alteration in obese youth.

Relationship between FGF-21 and whole-body, hepatic, and adipocyte insulin resistance (Figure 3)

Figure 3.

Correlation between FGF-21 and insulin sensitivity (glucose disposal, M), hepatic insulin resistance, and adipose insulin resistance. LBM, lean body mass.

The insulin-stimulated rate of glucose disposal (R_d) correlated inversely with plasma FGF-21 concentration (r² = −0.078, P = .004). A positive correlation also was observed between FGF-21 level and hepatic insulin resistance index (r² = 0.123, P < .001) and adipocyte insulin resistance index (r² = 0.051, P = .001). These correlations however were no longer significant when adjusting for hepatic fat content in a multiple stepwise analysis, suggesting that there is no independent effect of FGF-21 on peripheral and hepatic insulin sensitivity but rather the weak relationships maybe totally mediated by the coexistence of fatty liver disease.

Discussion

In this study we showed that, compared to healthy nonobese peers, obese youth have significantly increased circulating plasma FGF-21 levels. In addition to obesity, however, those with increased liver fat content displayed a further elevation in FGF-21 levels. The novel finding is, however, the independent relationships between FGF-21 not only with liver fat content but also with a marker of hepatic apoptosis, CK18 levels. In contrast, weak relationships between FGF-21, peripheral, hepatic, and adipocyte insulin sensitivity were found across the groups, which disappeared after adjusting for relevant confounders such as hepatic fat content.

Previous studies in adults reported an association between FGF-21 levels and indirect measure of adiposity, such as BMI values (24–28). Such a similar association has been also recently reported in youth by Reinehr et al (11). In our study we confirm an association between both BMI and BMI-Z with circulating FGF-21. In addition, we also described a significant association between FGF-21 concentrations and direct measures adiposity and adipose tissue distribution, evaluated by using the total body composition (dual-energy x-ray absorptiometry) and abdominal MRI.

The FGF “endocrine” subgroup, and especially FGF-21, has also been shown to be key factors in glucose and lipid metabolism and energy regulation (3, 4). In a group of adult patients with newly diagnosed type 2 diabetes, Chen et al reported significantly higher plasma FGF-21 concentrations compared to nondiabetic control subjects, and that FGF-21 negatively correlated with fasting plasma glucose (29). In a later study, Zhang et al (16) documented that circulating levels of FGF-21 correlated positively with fasting insulin and negatively with high-density lipoprotein cholesterol (26). Conversely, in patients with anorexia nervosa, plasma FGF-21 concentrations are decreased, and increased following weight gain (30). In contrast with these results, a recent study by Reinehr et al failed to show a similar association between FGF-21 and homeostasis model of assessment-insulin resistance in children and adolescents. However, in that study only fasting indexes of insulin sensitivity were used (11). In the current study we used not only indirect but also more direct and robust measures of insulin sensitivity to assess any potential relationship with FGF-21 in lean and obese adolescents. Indirect relationships were found with peripheral insulin-stimulated glucose disposal and direct relationships with both hepatic and adipocyte insulin sensitivity. These relationships were weak, however, and not independent of hepatic fat content.

Because the liver is the major site for FGF-21 expression and hepatic steatosis is highly correlated with impairment of glucose and lipid metabolism in humans, the relationship between hepatic steatosis and FGF-21 has been investigated in several recent studies. Li et al (10) reported that serum FGF-21 levels were significantly higher in the NAFLD group compared with the controls and had a high positive correlation with intrahepatic triglyceride content (r = 0.662, P < .001). This study, along with recent reports by Dushay et al (31) and Yilmaz et al (32), contributed greatly to expand our knowledge on plasma FGF-21 levels in patients with NAFLD and indicates the role of FGF-21 in regulating hepatic lipid metabolism. Only one recent study by Reinehr et al also reported such an association between ultrasound-detected hepatic fat content and FGF-21 levels (11). Although the aforementioned studies suggest that FGF-21 could be a potential biomarker to screen or monitor NAFLD patients (33), the methods used to assess the severity of hepatic steatosis, such as B-mode ultrasound or pathological score system, were qualitative or semi-quantitative and did not reflect the quantitative association between serum FGF-21 and hepatic fat content accurately. Moreover, in the study by Li et al, liver biopsies were obtained from patients undergoing resection for benign liver disease and the number of patients with precise information of hepatic fat content was rather small, which might preclude a reliable conclusion (10). In a recent study in adults using ¹H magnetic resonance spectroscopy to quantify hepatic fat content, Yan et al (34) showed that FGF-21 was strongly correlated with the hepatic fat content in people with mild or moderate hepatic steatosis and could better reflect hepatic fat content than any known serum parameters. In the present study, by using MRI and biopsy-derived measures of hepatic fat content and apoptosis, we documented a relevant correlation between FGF-21 and obesity-related liver alteration. In addition, by using multiple stepwise linear regression analysis and testing the variables of interest singularly in the model, we documented that FGF-21, BMI-Z, WBISI, and visceral fat were significantly associated with both hepatic fat content and CK18 levels, a index of hepatic apoptosis. More importantly, when all the variables were included in the model, FGF-21 levels were shown to be the best predictor. Finally, FGF-21 was directly and significantly correlated with the NAS score in the subgroup of youth who performed a liver biopsy.

The strength of the current study is a result of a relatively large sample size and the availability of both indirect and direct measures of adiposity, insulin sensitivity, and especially of liver hepatic fat content. Limitations are mainly related to the fact that we used a clinic-based sample of obese youths, which may not be a true representative of the general population. In addition, although our data suggest that FGFs could represent a possible marker of NAFLD in obese youth, due to the study design, the cause and effect relationships cannot be evaluated by the present cross-sectional clinical study. In addition, although we documented a correlation between FGF-21 values and NAS score, it should be acknowledged that in our study a persistent elevation in ALT level of the population was documented in 14 subjects and therefore a relatively small sample size of young subjects underwent liver biopsy. We acknowledge also that CK18 might not be specific for NASH because it can be elevated in a number of other conditions and thus may sometimes give spurious results (35). In fact, it has been shown that one possible limitation of the use of CK18 for the prediction of NASH is its intrinsic inability to discriminate between NASH and other chronic diseases that involve apoptosis, such as cholangitis and cholestasis, chronic hepatitis, cancer, and trauma (35).

Conclusions

Compared to healthy nonobese peers, circulating FGF-21 levels are increased in obese youth and are associated with fatty liver and markers of hepatic apoptosis. These associations are independent of overall obesity, visceral fat, and insulin sensitivity.