Maggio, Albane B.R.*; Mueller, Pascal†; Wacker, Julie‡; Viallon, Magalie§; Belli, Dominique C.||; Beghetti, Maurice‡; Farpour-Lambert, Nathalie J.*; McLin, Valérie A.||
The prevalence of childhood obesity is increasing rapidly in developing countries, resulting in the increased prevalence of associated comorbidities such as dyslipidemia, insulin resistance, cardiovascular diseases, sleep apnea, or liver disease. Twenty percent of Swiss children and adolescents are considered to be overweight and 5% to 8% of children are obese (1), mirroring what is known in other Western countries.
Obesity is described as body fat excess. It has been shown that fat distribution is central to the development of complications (2). Many organs, including the pancreas and the liver, are involved (3); however, the clinical implication of fat infiltration in the pancreas is not so clear. In rodents, pancreatic fat accumulation is related to obesity, triglycerides (TG), and increased cytokines (4). In humans, pancreatic fat accumulation has been shown in autopsy specimens to be related to age, overweight (5), and type 2 diabetes mellitus (T2DM) (6). An association with age and weight but not with T2DM was confirmed using computed tomography imaging in children and adults as well as on adult autopsy specimens (7); however, because there are few adult studies and only 1 in adolescents (8), the contribution of pancreatic fat to the development of insulin resistance and T2DM (9–11) remains unclear. Many studies have shown that metabolic derangements and insulin resistance secondary to central obesity contribute to nonalcoholic fatty liver disease and nonalcoholic steatohepatitis (12), but to our knowledge, no study to date has examined the relation between pancreatic fat and metabolic syndrome (MetS), although the tools are available.
Indeed, magnetic resonance imaging (MRI) has been shown to be a good tool to quantify both pancreatic and liver fat. The phase-shift imaging (modified Dixon method) is a rapid, reproducible, and operator-independent technique that is able to calculate the hepatic fat fraction (HFF) and the pancreatic fat fraction (PFF) (13,14). This method of PFF quantification has been validated against magnetic resonance spectroscopy (13). The aims of this project were to compare MRI-calculated PFF in obese and lean adolescents and to explore its relation to MetS, glucose and insulin levels during an oral glucose tolerance test (OGTT), biological markers of inflammation, and HFF.
Study Design and Subjects
The present study was nested in a prospective cross-sectional protocol aiming to measure cardiovascular risk factors in obese adolescents. The study population was composed of 50 obese and lean adolescents, ages 10 to 16 years. Obese subjects were recruited from the obesity clinic at our institution. Obesity was defined as body mass index (BMI) above the 97th age- and sex-specific percentiles (15). Lean children were recruited through advertising. Their height and weight were within the normal range (±2 standard deviation) and their BMI was <90th percentile for age and sex.
Inclusion criteria were determined for the main study. These were onset of puberty, no previous diagnosis of hypertension, no previous antihypertensive or antidiabetic treatment, no previous use of hepatotoxic drugs or medications affecting glucose and lipid metabolism, no history of familial hypertension or dyslipidemia, no known chronic hepatitis, and absence of diabetes or other chronic disease (eg, Wilson disease). Patients who were diagnosed as having hypertension during the enrollment period were not excluded if the previous criteria were not met; in other words, in the absence of a family history or medical history of hypertension, elevated blood pressure measurements at enrollment were considered to be a comorbidity associated with overweight/obesity. We did not exclude patients with a history of prematurity, small for gestational age, or intrauterine growth retardation. One obese adolescent was excluded after initial laboratory testing because of previously unknown familial hypercholesterolemia. A total of 24 obese and 25 lean adolescents completed the study.
Informed, written consent was obtained from both parents and children. The mother and child ethics committee of the University Hospitals of Geneva approved the study.
We assessed body weight (kilogram), height (centimeter), and hip and waist circumference (centimeter) using a nonelastic flexible tape. BMI was calculated as weight/height squared (kg · m−2) and z scores were derived using World Health Organization references (16). The determination of pubertal development (Tanner stages) was self-reported and photo recognition was also used.
Lipids and Hepatic Markers
Blood samples were collected by phlebotomy following a 10-hour overnight fast. Fasting total cholesterol, high-density lipoprotein cholesterol (HDL-C), TG, γ-glutamyltransferase (GGT), aspartate aminotransferase, and alanine aminotransferase (ALT) levels (units per liter) were determined using standard automated techniques (SYNCHRON LX20; Beckman Coulter, Brea, CA). Low-density lipoprotein cholesterol level (units per liter) was calculated according to the Friedewald formula.
Fasting plasma glucose and insulin concentration were measured using standard automated techniques (SYNCHRON LX20) and radioimmunoassay (Access ultrasensitive insulin; Beckman Coulter), respectively. Insulin resistance was assessed using the homeostasis model (HOMA-IR = fasting insulin [microunits per liter] × fasting glucose [millimoles per liter]/22.5). HOMA-IR >3 was considered abnormal (17).
Obese adolescents underwent an OGTT in the morning (8:00 AM) after a 12-hour overnight fast. Timed blood samples (at −15, 0, 30, 60, 90, 120, and 180 minutes) were collected for the measurement of plasma glucose and insulin concentrations (18).
Impaired fasting glucose, impaired glucose tolerance, or T2DM was defined according to the American Diabetes Association's guidelines (19). The insulogenic index was calculated as the ratio of the increment in plasma insulin level to that of plasma glucose level during the first 30 minutes after glucose ingestion (20).
Biomarkers and Inflammatory Markers
Serum leptin, adiponectin, chemokine ligand 2 (CCL2), interleukin-6 (IL-6), and tumor necrosis factor (TNF)-α levels were measured by colorimetric enzyme-linked immunosorbent assay (R&D Systems, Minneapolis, MN), following the manufacturer's instructions. The limits of detection were the following: leptin 31.25 pg/mL, adiponectin 62.50 pg/mL, CCL2 15.625 pg/mL, IL-6 0.156 pg/mL, and TNF-α 15.625 pg/mL. Mean intra- and interassay coefficients of variation were <7% for all of the markers.
Systemic blood pressure was measured 3 times, at 2-minute intervals, and after 10 minutes of rest in the supine position, using a standard, automated device (Colin Press-Mate BP 8800C, Colin Medical, San Antonio, TX). Hypertension was defined as blood pressure above the 95th sex-, age-, and height-specific percentiles (21).
The MetS was defined according to the International Diabetes Federation consensus for children and adolescents (22). It included increased waist circumference with increased fasting glucose or glucose intolerance, increased blood pressure, TG, or reduced HDL-C level.
Pancreatic and Hepatic Fat Quantification
Pancreatic and hepatic fat were quantified from MRI datasets. Patients were scanned in the supine position using a 1.5-T scanner (MAGNETOM Avanto System; Siemens HealthCare, Erlangen, Germany) equipped with a multiphase-array surface coil. PFF and HFF were obtained using the Dixon technique (23), a method of separating fat and water signals. Several steps were used to cover the whole abdomen from the diaphragmatic hepatic surface to the L5 vertebra. Water and fat image reconstruction from the acquired multiecho datasets was performed online (Syngo Software, Siemens HealthCare) using a 3-echo 2-point Dixon approach enabling voxel-wise correction of T2* decay (24). The VIBE imaging sequence (a 3-dimensional gradient echo volume interpolated breathold sequence) (25) outputs a fat percentage map, that is, fat fraction based on the separated fat and water images. These images are thereafter denoted water, fat, sum, water fraction, and fat fraction images, respectively. Quantitative analysis was performed on a clinical postprocessing workstation (MV-1000, Siemens Medical). Measurements of HFF and PFF were performed throughout the liver, using 15 regions of interest (ROI) (5 per slice, 3 slices per subject), and in the head and tail of the pancreas using 9 ROIs (3 per slice, 3 slices per subject). ROI selection avoided hepatic and pancreatic vessels, motion artifacts, or partial volume effects. Mean values of HFF and PFF were calculated from the previously defined ROIs. HFF >9% was considered to be steatosis (14). Intraobserver variability for PFF in a subset of 12 subjects was 3% (2.3% in 6 lean and 6.4% in 6 obese subjects). (Supplementary Methods are available online only at http://links.lww.com/MPG/A79.)
Visceral Fat Quantification
Total (TAAT), visceral (VAT), and subcutaneous abdominal adipose tissue (SAAT) were determined based on previously described sequences and postprocessed fat and water maps (2,26) obtained using the same 1.5-T MRI system (MAGNETOM Avanto System; Siemens HealthCare). To reduce interobserver variability, all of the images were analyzed offline by the same experienced observer, and postprocessing was performed on an Osirix workstation (Osirix Foundation, Geneva, Switzerland). The volume of total adipose tissue was obtained using the diaphragmatic hepatic surface as the upper limit and the iliac crests as the lower limit.
Statistical analyses were performed using SPSS software 15.0 (SPSS Inc, Chicago, IL). Data were screened initially for normalcy, using skewness and kurtosis tests. Several variables were transformed and successfully normalized: insulin (log10), GGT (log10), and TG (log10). We used nonparametric tests (Mann-Whitney) for variables that failed to normalize: leptin, CCL-2, IL-6, TNF-α, and HFF. Data are presented as mean and standard deviation scores or median and interquartile range when indicated. Statistical differences between groups were analyzed using independent Student t test and χ2 or Fisher exact test for normally distributed variables and analysis of variance and covariance. We evaluated the relation between dependent and independent variables using univariate and multivariate regression analyses. We excluded cases only if data were missing for the specific analysis. Differences were considered significant if P < 0.05.
Patients’ characteristics and fat distribution are presented in Table 1. As expected from the study design, obese subjects had significantly higher BMI, waist and hip circumferences, and visceral and subcutaneous fat than did lean subjects. Systolic blood pressure was also higher in the obese group. In each group, anthropometric variables and PFF were similar in both sex, except subcutaneous abdominal fat, which was significantly higher in obese girls (41.8 ± 6.8 vs 35.0 ± 7.7 cm3; P = 0.008) compared with obese boys. There was no relation between PFF and age in our series (r = 0.196, P = 0.177).
The results of blood markers are presented in Table 2 and show significant differences between groups for leptin concentration and for the components of MetS. Surprisingly, fasting glucose was lower in obese subjects. Six of 24 (25%) obese subjects presented with MetS and none in the lean group. In the obese group, the MetS was not related to sex or fat distribution (P > 0.05 for all).
High-sensitivity C-reactive protein was significantly different among groups, whereas other inflammatory markers were not (Table 2).
PFF Comparison Between Groups
PFF was significantly different between groups (Table 1). Because no official cutoff value exists for PFF, we divided our data in quartiles and created 2 categories: above or under the 75th percentile (cutoff point at 5.0%). With this cutoff, 37 subjects were in the normal PFF category (23/25 lean and 14/24 obese adolescents; mean PFF 3.6% ± 0.7%) and 12 in the elevated PFF category (2/25 lean and 10/24 obese adolescents, khi2: P = 0.006; mean PFF 5.9% ± 0.7%, P < 0.001). With this cutoff, we found that adiposity, HDL-C, and leptin levels were significantly different between the 2 PFF categories. This cutoff for PFF appears to correlate with both clinical and biochemical markers, making it clinically useful.
Associations Between PFF, Biochemical Markers, and Visceral Fat
Using univariate analysis, PFF was related to adiposity, MetS, GGT, TG, HDL-C, and leptin levels (Table 3). Surprisingly, when considering all of the subjects, insulin level was not related to PFF. In fact, it seems that in a subgroup analysis, the regression slope is positive in lean controls (t = 1.6, P = 0.114) but negative in obese subjects (t = −1.8, P = 0.087), confirming the absence of link between insulin and PFF.
To identify independent contributors of PFF, we selected the 5 following noncollinear covariates to perform multiple regression analysis: VAT, GGT, TG, HDL-C, and leptin levels. Results are summarized in Table 4; visceral fat was the only contributor to PFF, and there was a trend for HDL-C. When adjusted for visceral fat, PFF remained different between obese and lean subjects (F = 13.45, P < 0.001) and was related to the MetS in obese subjects (F = 13.4, P < 0.001).
To further investigate the relation between PFF and the MetS, we divided the subjects into 3 groups: group I, lean without MetS; group II, obese without MetS; and group III, obese with MetS. Results showed that PFF increased gradually between groups (PFF: group I [n = 25] 3.56 ± 0.88; group II [n = 19] 4.70 ± 1.06; group III [n = 5] 5.34 ± 1.49%; F = 10.36, P < 0.001) even after adjustment for VAT (F = 9.2, P < 0.001) (Fig. 1); however, PFF was not significantly different between obese subjects with or without MetS (P > 0.05). Mean VAT was different between group I (5.0 ± 2.4) and the 2 other groups (group II: 12.3 ± 4.8; group III: 14.7 ± 3.9 cm3; P < 0.05 for both), but not between the 2 obese groups (groups II and III: P = 0.307).
To confirm this finding, we looked for factors associated with PFF in obese adolescents only. We found a significant association between PFF and HDL-C concentrations (P = 0.025), but not with BMI or fat distribution. There was no association with OGTT results; however, when PFF was adjusted for VAT, the following OGTT measures were inversely related to PFF: fasting insulin, 30-minute, 120-minute, and maximum insulin levels (fasting: R2 change 0.223, P = 0.016, β = −0.540, t = −2.62; 30-minute: R2 change 0.253, P = 0.014, β = −0.545, t = −2.69; 120-minute: R2 change 0.202, P = 0.031, β = −0.461, t = −2.32; maximum: R2 change 0.288, P = 0.008, β −0.607, t −2.94). There was no association with serum glucose.
Among the inflammatory markers analyzed using a univariate regression, PFF was only related to serum CCL-2 concentrations (Table 3). There was no correlation between inflammatory markers and glucose or insulin levels at any time during the OGGT (P > 0.05 for all).
Hepatic Fat Fraction
HFF was significantly higher in obese subjects compared with lean subjects (Table 1) and was correlated with BMI, visceral and subcutaneous fat, alanine aminotransferase, GGT, lipids, insulin, and leptin levels (P < 0.05). Surprisingly, it was not related to MetS (t = 1.33, P = 0.189) in this series.
In obese children, HFF correlated with OGTT 60-minute glucose (t = 3.4, P = 0.003) and insulin (t = 3.8, P = 0.001); however, fasting glucose (t = 0.016, P = 0.987), fasting insulin (t = 1.9, P = 0.072), and the insulinogenic index (t = −0.45, P = 0.658) did not correlate. To mirror our PFF analysis, we adjusted HFF for VAT. This analysis confirmed our findings, and in addition, the following 2 parameters correlated with HFF: 120-minute glucose (R2 change 0.145, P = 0.035) and maximal insulin secretion (R2 change 0.194, P = 0.016).
Association Between PFF and HFF
Table 3 illustrates that there was no linear correlation between PFF and HFF, even when analyzing the groups separately (lean: t = 1.57, P = 0.130; obese: t = −0.043, P = 0.841). Furthermore, PFF distribution was wide, whereas HFF values clustered within a narrow range (Fig. 2). Only 2 obese adolescents had both increased PFF and increased HFF.
Our study demonstrates that obese adolescents have greater pancreatic fat accumulation than lean controls. For the first time, increased PFF is shown to be associated with MetS, visceral fat, and impaired insulin response to glucose overload in obese subjects. Based on these data, we propose that the upper limit of normal for pancreatic fat content, by MRI, in this age group is 5% of total pancreatic volume. It is notable that PFF was not significantly associated with HFF in this population, suggesting different pathophysiology.
Increased PFF in Obese Adolescents
The present study illustrates that obese adolescent boys and girls (mean age 13 years) have increased pancreatic fat deposition compared with lean controls, confirming findings in adolescent girls (8). Unlike what has been published for the liver, no cutoff values have been published for PFF quantification using MRI. One adult study using MRI spectroscopy evaluated normal pancreatic fat content to be <2.8 fat/water% (11). To this end, we propose a 5% cutoff value for our population, which appears to adequately reflect fat accumulation in the gland itself, independent of visceral adiposity. This novel cutoff remains to be confirmed in other lean and obese populations.
PFF Is Associated With MetS in Obese Adolescents
MetS has been shown to be strongly associated with elevated HFF and hepatic steatosis in adults and adolescents alike (12,27). In our study, we show that in a small but representative cohort of obese adolescents, PFF was also related to the presence of MetS. In other words, in obese subjects with a PFF>5%, there may be an increased risk of developing MetS.
To identify predictors of developing MetS, we looked at different factors implicated in MetS and in fat deposition. We found that BMI, subcutaneous fat, visceral fat, leptin, systolic blood pressure, and lipids were associated with PFF. Lipids, especially HDL-C and TG, were involved, but visceral fat was the strongest predictor of PFF.
How PFF and MetS are related remains to be elucidated. We observed that PFF increased as BMI increased both in lean controls and in obese subjects with and without MetS. The fact that there was no difference in VAT between obese adolescents with or without MetS suggests that the trend for increasing PFF is not the result of higher VAT but rather of another mechanism. Based on our results, HDL-C rather than TG or visceral fat may be related to PFF; however, the association between PFF and HDL-C in obese adolescents may be simply explained by the fact that low HDL-C and high PFF are present concomitantly in the most severely affected individuals without a clear causality (28).
Elevated PFF Predicts Impaired Insulin Secretion in Obese Adolescents
Moreover, we tried to evaluate the relation between elevated PFF and glucose metabolism in obese adolescents. Despite normal, mean fasting insulin concentration, obese adolescents had significantly higher fasting insulin levels than did lean controls. The OGGT showed no impaired glucose tolerance; however, when PFF was adjusted for visceral fat, insulin levels were negatively related to PFF, whereas this relation was positive in lean controls. In fact, patients with higher pancreatic fat deposition had lower insulin secretion in response to glucose overload, suggestive of early β-cell dysfunction.
Whether fat accumulation within the pancreas is a cause or consequence of β-cell dysfunction is not clear; however, evidence from the literature would support the former. First, it has been shown in humans that β-cell damage is present for a long period of time before β-cell dysfunction and the onset of diabetes (29). Second, in a study comparing age- and BMI-matched adults with and without T2DM, pancreatic fat content was related to β-cell function parameters (including insulogenic index and HOMA-IR) only in nondiabetic subjects, suggesting that pancreatic fat accumulation may precede the onset of T2DM (9). Furthermore, they found that pancreatic fat content was higher in T2DM individuals compared with normoglycemic BMI-matched subjects, suggesting either pancreatic fat continues to accumulate as diabetes progresses or the advent of T2DM leads to an increase in pancreatic fat accumulation. The association between pancreatic lipomatosis and T2DM was also confirmed in autopsy specimens (6), but this finding does not help in establishing causality (30–32).
Serum CCL-2 Is Correlated With PFF
Adipose tissue is a metabolically active tissue and mediates some of its effects via potent adipokines and cytokines (33–35). It is unclear where abnormal fat accumulates in the pancreas and how it may contribute to β-cell dysfunction. Based on the scarce data available, it is possible that fat accumulates in the mesenchyme surrounding acinar tissue and exerts a paracrine effect on neighboring β-cells via cytokines, which in turn leads to β-cell dysfunction (36).
TNF-α and IL-6, 2 potent mediators of inflammation, are expressed in adipose tissue and implicated in the development of insulin resistance (33,35,37). Serum levels of these 2 markers were not significantly different between obese and lean adolescents, and they were not associated with PFF or insulin resistance; however, C-reactive protein, also known to be both increased in obese subjects and related to insulin resistance (38), was elevated in our obese patients, albeit not related to insulin resistance or PFF. Finally, CCL-2 is known to be increased in obese subjects and linked to insulin resistance and T2DM in adults (34,39). Although serum CCL-2 concentrations were not significantly different between lean and obese groups, it did correlate with PFF in our univariate analysis, but not with insulin resistance. In summary, although the relation between lipid accumulation and inflammation is now well accepted in the pathogenesis of obesity-related complications, our study was unable to demonstrate a strong link between conventionally accepted inflammatory mediators and PFF or insulin resistance. We ascribe this either to the cohort's relatively young mean age (fewer obesity-related complications) or to the small sample size.
No Significant Correlation Between PFF and HFF
Previous studies have shown that PFF and HFF are either weakly or not at all correlated (9,11,13,40). We confirm this impression in our small but representative cohort. In children, HFF is related to BMI, visceral and subcutaneous fat, and serum lipids (8). Our results also confirm that HFF is related to insulin metabolism in obese adolescents, but not to PFF (41).
The distribution of lipid accumulation in the pancreas of our 49 study participants is markedly different from that seen in the liver. Although PFF shows a wide distribution, HFF shows a comparatively narrow distribution. This finding is intriguing and suggests a different timing and underlying mechanism leading to fat accumulation in each organ.
In conclusion, obese adolescents have higher pancreatic and hepatic fat deposition than do lean controls, by MRI, but the 2 are not related. Obese adolescents with MetS appear to have higher PFF than their peers without MetS. Furthermore, PFF was associated with visceral fat distribution and low serum HDL-C concentration, something previously unreported. Moreover, these data suggest that the effect of pancreatic fat deposition is associated with an impaired response to glucose overload in obese adolescents, suggestive of early β-cell dysfunction. Further studies are needed in children and adolescents to characterize the association between pancreatic fat accumulation and the development of MetS and T2DM.
We thank the subjects for volunteering for the study and the staff of the Pediatric Research Platform. The authors thank the Cardiology Division of the Foundation for Medical Research for technical support; W.D. Gilson and S. Kannengiesser, Siemens Healthcare, for providing the “VIBE with T2*-corrected Dixon Fat/Water Separation” package; and Stephanie H. Abrams, MD, for helpful discussion.
1. Zimmermann MB, Gubeli C, Puntener C, et al. Overweight and obesity in 6-12 year old children in Switzerland. Swiss Med Wkly 2004; 134:523–528.
2. Gutin B, Johnson MH, Humphries MC, et al. Relationship of visceral adiposity to cardiovascular disease risk factors in black and white teens. Obesity (Silver Spring) 2007; 15:1029–1035.
3. van Herpen NA, Schrauwen-Hinderling VB. Lipid accumulation in non-adipose tissue and lipotoxicity. Physiol Behav 2008; 94:231–241.
4. Mathur A, Marine M, Lu D, et al. Nonalcoholic fatty pancreas disease. HPB (Oxford) 2007; 9:312–318.
5. Olsen TS. Lipomatosis of the pancreas in autopsy material and its relation to age and overweight. Acta Pathol Microbiol Scand A 1978; 86A:367–373.
6. Stamm BH. Incidence and diagnostic significance of minor pathologic changes in the adult pancreas at autopsy: a systematic study of 112 autopsies in patients without known pancreatic disease. Hum Pathol 1984; 15:677–683.
7. Saisho Y, Butler AE, Meier JJ, et al. Pancreas volumes in humans from birth to age one hundred taking into account sex, obesity, and presence of type-2 diabetes. Clin Anat 2007; 20:933–942.
8. Kovanlikaya A, Mittelman SD, Ward A, et al. Obesity and fat quantification in lean tissues using three-point Dixon MR imaging. Pediatr Radiol 2005; 35:601–607.
9. Tushuizen ME, Bunck MC, Pouwels PJ, et al. Pancreatic fat content and beta-cell function in men with and without type 2 diabetes. Diabetes Care 2007; 30:2916–2921.
10. Carpentier AC, Bourbonnais A, Frisch F, et al. Plasma nonesterified fatty acid intolerance and hyperglycemia are associated with intravenous lipid-induced impairment of insulin sensitivity and disposition index. J Clin Endocrinol Metab 2010; 95:1256–1264.
11. Lingvay I, Esser V, Legendre JL, et al. Noninvasive quantification of pancreatic fat in humans. J Clin Endocrinol Metab 2009; 94:4070–4076.
12. Manco M, Marcellini M, Devito R, et al. Metabolic syndrome and liver histology in paediatric non-alcoholic steatohepatitis. Int J Obes (Lond) 2008; 32:381–387.
13. Schwenzer NF, Machann J, Martirosian P, et al. Quantification of pancreatic lipomatosis and liver steatosis by MRI: comparison of in/opposed-phase and spectral-spatial excitation techniques. Invest Radiol 2008; 43:330–337.
14. Fishbein MH, Gardner KG, Potter CJ, et al. Introduction of fast MR imaging in the assessment of hepatic steatosis. Magn Reson Imaging 1997; 15:287–293.
15. Rolland-Cachera MF, Cole TJ, Sempe M, et al. Body mass index variations: centiles from birth to 87 years. Eur J Clin Nutr 1991; 45:13–21.
16. WHO Child Growth Standards: Length/Height-for-Age, Weight-for-Age, Weight-for-Length, Weight-for-Height and Body Mass Index-for-Age: Methods and Development. Geneva: World Health Organization; 2006:312.
17. Allard P, Delvin EE, Paradis G, et al. Distribution of fasting plasma insulin, free fatty acids, and glucose concentrations and of homeostasis model assessment of insulin resistance in a representative sample of Quebec children and adolescents. Clin Chem 2003; 49:644–649.
18. Matsuda M, DeFronzo RA. Insulin sensitivity indices obtained from oral glucose tolerance testing: comparison with the euglycemic insulin clamp. Diabetes Care 1999; 22:1462–1470.
19. American Diabetes Association. Diagnosis and classification of diabetes mellitus. Diabetes Care 2006;29(suppl 1):S43–8.
20. Phillips DI, Clark PM, Hales CN, et al. Understanding oral glucose tolerance: comparison of glucose or insulin measurements during the oral glucose tolerance test with specific measurements of insulin resistance and insulin secretion. Diabet Med 1994; 11:286–292.
21. National High Blood Pressure Education Program Working Group on High Blood Pressure in Children and Adolescents. The fourth report on the diagnosis, evaluation, and treatment of high blood pressure in children and adolescents. Pediatrics 2004;114:555–76.
22. Jolliffe CJ, Janssen I. Development of age-specific adolescent metabolic syndrome criteria that are linked to the Adult Treatment Panel III and International Diabetes Federation criteria. J Am Coll Cardiol 2007; 49:891–898.
23. Dixon WT. Simple proton spectroscopic imaging. Radiology 1984; 153:189–194.
24. Hines CD, Yu H, Shimakawa A, et al. T1 independent, T2* corrected MRI with accurate spectral modeling for quantification of fat: validation in a fat-water-SPIO phantom. J Magn Reson Imaging 2009; 30:1215–1222.
25. Rofsky NM, Lee VS, Laub G, et al. Abdominal MR imaging with a volumetric interpolated breath-hold examination. Radiology 1999; 212:876–884.
26. Brambilla P, Bedogni G, Moreno LA, et al. Crossvalidation of anthropometry against magnetic resonance imaging for the assessment of visceral and subcutaneous adipose tissue in children. Int J Obes (Lond) 2006; 30:23–30.
27. D’Adamo E, Marcovecchio ML, Giannini C, et al. The possible role of liver steatosis in defining metabolic syndrome in prepubertal children. Metabolism 2010; 59:671–676.
28. Barter P, Gotto AM, LaRosa JC, et al. HDL cholesterol, very low levels of LDL cholesterol, and cardiovascular events. N Engl J Med 2007; 357:1301–1310.
29. Overview of 6 years’ therapy of type II diabetes: a progressive disease. U.K. Prospective Diabetes Study Group. Diabetes 1995;44:1249–58.
30. Lee Y, Lingvay I, Szczepaniak LS, et al. Pancreatic steatosis: harbinger of type 2 diabetes in obese rodents. Int J Obes (Lond) 2010; 34:396–400.
31. McGarry JD. Banting lecture 2001: dysregulation of fatty acid metabolism in the etiology of type 2 diabetes. Diabetes 2002; 51:7–18.
32. El-Assaad W, Buteau J, Peyot ML, et al. Saturated fatty acids synergize with elevated glucose to cause pancreatic beta-cell death. Endocrinology 2003; 144:4154–4163.
33. Rotter V, Nagaev I, Smith U. Interleukin-6 (IL-6) induces insulin resistance in 3T3-L1 adipocytes and is, like IL-8 and tumor necrosis factor-alpha, overexpressed in human fat cells from insulin-resistant subjects. J Biol Chem 2003; 278:45777–45784.
34. Sartipy P, Loskutoff DJ. Monocyte chemoattractant protein 1 in obesity and insulin resistance. Proc Natl Acad Sci U S A 2003; 100:7265–7270.
35. Hotamisligil GS, Spiegelman BM. Tumor necrosis factor alpha: a key component of the obesity-diabetes link. Diabetes 1994; 43:1271–1278.
36. Mandrup-Poulsen T. Beta-cell apoptosis: stimuli and signaling. Diabetes 2001; 50 (suppl 1):S58–S63.
37. Bastard JP, Maachi M, Van Nhieu JT, et al. Adipose tissue IL-6 content correlates with resistance to insulin activation of glucose uptake both in vivo and in vitro. J Clin Endocrinol Metab 2002; 87:2084–2089.
38. Festa A, D’Agostino R Jr, Howard G, et al. Chronic subclinical inflammation as part of the insulin resistance syndrome: the Insulin Resistance Atherosclerosis Study (IRAS). Circulation 2000; 102:42–47.
39. Xu H, Barnes GT, Yang Q, et al. Chronic inflammation in fat plays a crucial role in the development of obesity-related insulin resistance. J Clin Invest 2003; 112:1821–1830.
40. van Geenen EJ, Smits MM, Schreuder TC, et al. Nonalcoholic fatty liver disease is related to nonalcoholic fatty pancreas disease. Pancreas 2010; 39:1185–1190.
41. Cali AM, De Oliveira AM, Kim H, et al. Glucose dysregulation and hepatic steatosis in obese adolescents: is there a link? Hepatology 2009; 49:1896–1903.