One-hour post-load plasma glucose levels are associated with elevated liver enzymes

Nutrition, Metabolism & Cardiovascular Diseases (2011) 21, 713e718

available at www.sciencedirect.com

journal homepage: www.elsevier.com/locate/nmcd

One-hour post-load plasma glucose levels are associated with elevated liver enzymes E. Succurro, F. Arturi, A. Grembiale, F. Iorio, T.V. Fiorentino, F. Andreozzi, A. Sciacqua, M.L. Hribal, F. Perticone, G. Sesti* Department of Experimental and Clinical Medicine, viale Europa, University Magna-Græcia of Catanzaro, Catanzaro 88100, Italy Received 24 August 2010; received in revised form 26 January 2011; accepted 3 February 2011

KEYWORDS Glucose tolerance; Liver enzymes; NAFLD

Abstract Background and aims: Glucose-tolerant subjects who have 1-h post-load glucose levels
155 mg dl1 (normal glucose tolerance (NGT)-1h-high) are at an increased risk of developing type 2 diabetes. Prospectively conducted studies indicated that high levels of liver enzymes are predictors of a tendency to develop type 2 diabetes; however, it is unknown whether the NGT-1h-high subjects are at increased risk for secreting higher levels of liver biomarkers. Methods and results: In this study, oral glucose tolerance tests (OGTTs) were performed in a cohort of 1000 non-diabetic Caucasians and levels of alanine aminotransferase (ALT), aspartate aminotransferase (AST) and gamma-glutamyltransferase (GGT) were measured in these subjects. The NGT-1h-high subjects had increased levels of ALT and GGT, but not AST, as compared with the NGT-1h-low. Following adjustment for age and gender, the ALT, AST and GGT levels were all found to be significantly correlated with body mass index (BMI), waist circumference, blood pressure, triglycerides as well as fasting and post-challenge glucose and insulin levels. In a logistic regression analysis adjusted for age and gender, NGT-1h-high subjects were found to be at increased risk of having ALT levels in the highest quartile as compared with NGT-1h-low subjects (odds ratio (OR) Z 1.71; 95% confidence interval (CI): 1.16e2.52). In addition, NGT-1 h-high subjects exhibited an increased risk for having GGT levels in the highest quartile (OR Z 1.50; 95%CI: 1.02e2.17). These associations remained significant after adjustment for BMI, blood pressure and lipids, but were not significant following further adjustment for an insulin sensitivity index. NGT-1h-high subjects were at increased risk of having AST levels in the highest quartile as compared with NGT-1h-low subjects (OR Z 1.51; 95%CI: 1.04e2.22). This association ceased to be significant following adjustment for BMI, blood pressure and lipids. Conclusions: These data suggest that a 1hPG
155 mg dl1 cut-off may facilitate the identification of NGT individuals at risk of developing liver abnormalities. ª 2011 Elsevier B.V. All rights reserved.

* Corresponding author. Tel.: þ39 0961 3647204; fax: þ39 0961 3647192. E-mail address: [email protected] (G. Sesti). 0939-4753/$ - see front matter ª 2011 Elsevier B.V. All rights reserved. doi:10.1016/j.numecd.2011.02.002

714 Subjects with impaired fasting glucose (IFG) and/or impaired glucose tolerance (IGT) are at high risk for the prospective development of type 2 diabetes. However, there is evidence that 30e40% of individuals who develop type 2 diabetes have normal glucose tolerance (NGT) at baseline [1]. In recent times, it has been demonstrated that a cut-off point, set at 155 mg dl1, for the 1-h post-load plasma glucose (PG) level during the oral glucose tolerance test (OGTT) can potentially identify subjects who are at relatively high risk of developing type 2 diabetes from among individuals who have NGT or IGT [2]. Moreover, subjects with NGT and a 1-h PG level
155 mg dl1 exhibit a poor cardio-metabolic risk profile as well as early signs of vascular atherosclerosis [3,4]. The liver plays an important role in maintaining glucose homeostasis during fasting as well as the postprandial state [5]. The term non-alcoholic fatty liver disease (NAFLD) refers to a spectrum of conditions that ranges from simple hepatic steatosis to more severe disorders, including nonalcoholic steatohepatitis (NASH), which can progress to fibrosis and cirrhosis. The NAFLD is associated with obesity, insulin resistance and type 2 diabetes [6,7], and is best diagnosed by liver biopsy or proton magnetic resonance spectroscopy e procedures that are not applicable in largescale epidemiological studies. The NAFLD is characterised by elevated levels of liver enzymes, including alanine aminotransferase (ALT), aspartate aminotransferase (AST) and gamma-glutamyltransferase (GGT), in the circulation, and, therefore, these enzymes e in particular, ALT e have been extensively used as reliable surrogate measures for NAFLD in epidemiological studies. Several prospective studies have shown that high levels of ALT and GGT are independent predictors of incident type 2 diabetes [8e10]. It remains undefined whether subjects with a 1-h post-load glucose level
155 mg dl1 are at increased risk of having elevated levels of liver enzymes. There is evidence that patients with type 2 diabetes and NAFLD are at increased risk of developing fibrosis and cirrhosis [11]. Treating the underlying risk factors for NAFLD is the best approach for preventing or delaying the adverse outcomes of the condition, and lifestyle-related interventions, including diet and exercise, in high-risk subjects have consistently demonstrated their efficacy in decreasing the incidence of type 2 diabetes [12,13] and in improving NAFLD [14e17] and, therefore, it would be important to identify individuals who are at increased risk for the above-mentioned clinical outcomes. This study aims to assess whether Caucasian subjects who have elevated levels of 1-h postload PG are at increased risk for having higher levels of liver biomarkers associated with NAFLD in the circulation.

Methods Subjects in this study consisted of 1000 Caucasian subjects participating to the CAtanzaro MEtabolic RIsk factors Study (CATAMERIS), an observational study for the identification of cardio-metabolic risk factors; the exclusion criteria have been described previously and comprised known history of diabetes mellitus, chronic gastrointestinal diseases or pancreatitis, history of any malignant disease or drug abuse, self-reported alcohol intake of three or more drinks

E. Succurro et al. per day and test positivity for antibodies to hepatitis C virus (HCV) or hepatitis B surface antigen (HBsAg) [3,18e20]. On the first day, following 12 h of fasting, subjects underwent complete anthropometrical evaluation, and a venous blood sample was drawn for laboratory determination of baseline values. On the second day, following a 12 h overnight fast, a 75-g OGTT was performed with sampling at 0, 30, 60 and 120 min for PG and insulin; the subjects were classified into two groups as NGT (fasting PG (FPG) < 100 mg dl1 and 2-h post-load PG level <140 mg dl1) or IGT (FPG < 126 mg dl1 and 2-h post-load PG level 140e199 mg dl1). The protocol for the present study was approved by the ethical committee, and informed written consent was obtained from all participants. The study was performed in accordance with the principles of the Declaration of Helsinki.

Analytical determinations The ALT and AST levels were measured using the a-ketoglutarate reaction and GGT levels with the L-gamma-glutamyl-3-carboxy-4-nitroanilide rate method (Roche, Basel, Switzerland). Glucose, triglyceride, total and high-density lipoprotein (HDL)-C concentrations were determined by enzymatic methods (Roche, Basel, Switzerland) and the plasma insulin concentration was quantified by a chemiluminescence-based assay (Immulite, Siemens, Italy).

Calculations Insulin sensitivity was estimated by using a previously validated index (insulin sensitivity index (ISI)) [21], which was calculated by the following formula: 10,000/square root of [fasting glucose (millimoles per litre)  fasting insulin (milliunits per litre)]  [mean glucose  mean insulin during OGTT]. Men and women were separately stratified, by gender, into quartiles according to their ALT, AST or GGT levels. The highest quartiles of liver enzymes were identified and are as follows: ALT Z
38 IU l1 for men, and
26 IU l1 for women; AST Z
26 IU l1 for men, and
23 IU l1 for women; and GGT Z
37 IU l1 for men, and
25 IU l1 for women. Men and women stratified in the highest quartiles were then classified as a single group who constituted the risk category in a logistic regression analysis. In addition, elevated ALT values were defined to be >30 IU l1 for men, and >19 IU l1 for women, as has been previously reported [22].

Statistical analysis Fasting levels of insulin and insulin levels during OGTT were naturally log-transformed for statistical analysis because of their skewed distribution. Continuous data were expressed as means  standard deviation (SD). Categorical variables were compared by the chi-square (c2) test. A general, linear model, with post hoc Bonferroni correction for multiple comparisons, was used to compare phenotypic differences between the groups. Partial correlation coefficients adjusted for age and gender were computed between variables. A multivariable logistic regression analysis was used to determine the association between the study groups and higher levels of liver enzymes. A P-value <0.05 was

1h-OGTT and liver enzymes

715

considered to be statistically significant. All analyses were performed using the Statistical Package for Social Sciences (SPSS) software programme, version 16.0, for Windows.

Results Among the 1000 subjects examined, 752 (75.2%) had NGT and 248 (24.8%) had IGT. Subjects with NGT were classified into two groups: 532 subjects with 1-h post-load PG level <155 mg dl1 (NGT-1h-low) and 220 individuals with 1-h post-load PG level
155 mg dl1 (NGT-1h-high). Table 1 presents the clinical characteristics and laboratory findings of the three study (2 NGT þ 1 IGT) groups. Significant differences among these three groups were observed with regard to gender (higher prevalence of men among subjects with NGT-1h-high and IGT, as compared with NGT-1h-low) and age (subjects with NGT-1h-high and IGT were older than those with NGT-1h-low); therefore, all analyses were adjusted for age and gender. NGT-1h-high individuals had a metabolic risk profile that was intermediate between that observed in NGT-1h-low and IGT subjects. As shown in Table 1, NGT-1h-high subjects exhibited significantly higher BMI, waist circumference, fasting and 2-h post-load PG, 1-h and 2-h post-challenge

Table 1

insulin levels as well as lower insulin sensitivity, which was assessed by the ISI index, as compared with NGT-1h-low subjects. In addition, the former had increased levels of ALT and GGT, but not AST, as compared with the latter. A greater proportion of NGT-1h-high individuals had IFG, as compared with NGT-1h-low subjects. There were no differences in age and gender that could be observed between NGT-1h-high individuals and subjects with IGT (Table 1). Subjects with IGT exhibited significantly higher BMI, waist circumference, systolic blood pressure (SBP), triglycerides, fasting, 1-h and 2-h post-challenge insulin levels as well as lower insulin sensitivity levels as compared with NGT-1h-high individuals. In addition, these subjects had increased levels of ALT and GGT, but not AST, as compared with NGT-1h-high individuals. Age- and gender-adjusted univariate correlations between liver enzymes and anthropometric and metabolic variables for the entire study group showed that ALT, AST and GGT were all significantly correlated with BMI, waist circumference, SBP, DBP, triglycerides, fasting, 1-h and 2-h post-challenge PG and insulin levels (Table 2). Levels of total and HDL cholesterol showed moderate statistically significant correlation with ALT, whereas insulin sensitivity, assessed by the ISI index, was negatively correlated with all the three liver biomarkers (Table 2).

Anthropometric and metabolic characteristics of the study subjects stratified according to glucose tolerance.

Variables

NGT with 1-h glucose NGT with 1-h glucose IGT (3) <155 mg dl1 (1) >155 mg dl1 (2)

P

n (Male/Female) Age (yr) BMI (kg m2) Waist circumference (cm) SBP (mmHg) DBP (mmHg) Fasting glucose (mg dl1) 1-h glucose (mg dl1) 2-h glucose (mg dl1) Fasting insulin (mU ml1) 1-h insulin (mU ml1) 2-h insulin (mU ml1) Total cholesterol (mg dl1) HDL (mg dl1) Triglycerides (mg dl1) ALT (UI l1) AST (UI l1) GGT (UI l1) ISI IFG

532 (202/330) 42  13 28.9  6.1 95  14

<0.0001 <0.0001 0.0002 <0.0001a <0.0001a <0.0001a <0.0001b 0.004b <0.0001b <0.0001b 0.04b <0.0001b

0.24 0.06a 0.03b 0.002b

137  17 84  10 99  13

0.003 0.15 <0.0001

0.39 0.65 <0.0001

0.05 0.99 <0.0001

0.002 0.16 0.001

     

36 18 9 75 108 43

<0.0001 <0.0001 <0.0001 <0.0001 <0.0001 0.88

<0.0001 <0.0001 0.26 <0.0001 <0.0001 0.99

<0.0001 <0.0001 <0.0001 0.003 <0.0001 0.99

<0.0001 <0.0001 0.001 0.07 <0.0001 0.99

48  14 158  89 29  15 23  8 33  31 52  29 110 (44.4%)

0.10 <0.0001 <0.0001 0.52 0.001 <0.0001 <0.0001

0.60 0.99 0.03 0.99 0.05 <0.0001 <0.0001

0.11 <0.0001 <0.0001 0.82 0.005 <0.0001 <0.0001

0.99 <0.0001 0.21 0.99 0.37 0.20 0.06

127  17 80  10 87  9 117 100 11 79 56 197

     

23 19 8 59 48 40

54  21 111  57 23  13 20  6 22  14 98  52 58 (10.9%)

220 (128/92) 49  12 30.2  5.6 99  12 130  16 81  10 95  10 180 114 12 111 84 203

     

22 17 6 65 65 37

50  12 124  64 27  14 22  7 30  25 65  35 79 (35.9%)

248 (130/118) 52  12 31.6  5.8 104  14

192 165 15 107 135 206

P 1 vs. 2 P 1 vs. 3 P 2 vs. 3

Data are means  SD. Fasting, 1-h and 2-h insulin were log transformed for statistical analysis, but values in the table represent a back transformation to the original scale. Comparisons between the three groups were performed using a general linear model with post hoc Bonferroni correction for multiple comparisons. P values refer to results after analyses with adjustment for age, gender, and BMI. Categorical variables were compared by c2 test. NGT Z normal glucose tolerance; IGT Z impaired glucose tolerance; SBP Z systolic blood pressure; DBP Z diastolic blood pressure; IS I Z insulin sensitivity index; IFG Z Impaired fasting glucose. a P values refer to results after analyses with adjustment for gender and BMI. b P values refer to results after analyses with adjustment for age, and gender.

716

E. Succurro et al.

Table 2 Age- and gender-adjusted univariate correlations between liver enzymes and anthropometric and metabolic variables.

BMI (kg m2) Waist (cm) SBP (mmHg) DBP (mmHg) Fasting glucose (mg dl1) 1-h glucose (mg dl1) 2-h glucose (mg dl1) Fasting insulin (mU ml1) 1-h insulin (mU ml1) 2-h insulin (mU ml1) Total cholesterol (mg dl1) HDL (mg dl1)) Triglycerides (mg dl1) ISI

ALT

P

AST

P

GGT

P

r

P

r

P

r

P

0.28 0.30 0.17 0.17 0.14 0.16 0.21 0.33 0.27 0.32 0.07 0.10 0.20 0.32

<0.0001 <0.0001 <0.0001 <0.0001 <0.0001 <0.0001 <0.0001 <0.0001 <0.0001 <0.0001 0.03 0.003 <0.0001 <0.0001

0.16 0.17 0.09 0.11 0.09 0.12 0.13 0.13 0.09 0.08 0.004 0.05 0.12 0.17

<0.0001 <0.0001 0.008 0.001 0.007 <0.0001 <0.0001 <0.0001 0.01 0.02 0.90 0.11 <0.0001 <0.0001

0.15 0.19 0.09 0.11 0.11 0.08 0.09 0.19 0.15 0.14 0.09 0.02 0.23 0.18

<0.0001 <0.0001 0.008 0.001 <0.0001 0.01 0.002 <0.0001 <0.0001 <0.0001 0.002 0.53 <0.0001 <0.0001

BMI Z Body Mass Index; SBP Z systolic blood pressure; DBP Z diastolic blood; ISI Z insulin sensitivity index.

A logistic regression model, adjusted for age and gender, was used to compare the risk of NGT-1h-high and IGT groups to have liver enzymes in the highest quartile as compared with the NGT-1h-low group (the reference category). NGT1h-high subjects had a 1.7-fold increased risk of having ALT level in the highest quartile as compared with the NGT-1hlow group (odds ratio (OR): 1.71; 95% confidence interval (95%CI): 1.16e2.52; P Z 0.006). Furthermore, an increased risk of having ALT levels in the highest quartile was observed in IGT subjects (OR: 2.5; 95%CI: 1.79e3.74; P < 0.0001). Following additional adjustment for a wide range of potential confounders, including BMI (or waist circumference), SBP, DBP, triglycerides, total cholesterol and HDL, it was found that NGT-1h-high subjects continued to be associated with an increased risk of having elevated ALT concentrations (OR: 1.59; 95%CI: 1.06e2.40; P Z 0.02). These associations were abrogated when the ISI index of insulin sensitivity was additionally added to the model (OR: 1.26; 95% CI: 0.79e2.00; P Z 0.32). When elevated ALT value was defined as being >30 IU l1 for men and >19 IU l1 for women, according to a previously reported cut-off point [23], NGT-1h-high subjects were found to be at a 1.5-fold increased risk of having elevated ALT concentrations (OR: 1.50; 95%CI: 1.07e2.10; P Z 0.01). A similar increased risk of having elevated ALT concentrations was observed in subjects with IGT (OR: 2.13; 95%CI: 1.53e2.96; P < 0.0001). Following additional adjustment for a wide range of potential confounders, including BMI (or waist circumference), SBP, DBP, triglycerides, total cholesterol and HDL, NGT-1h-high subjects continued to be associated with an increased risk of having elevated ALT concentration (OR: 1.30; 95%CI: 1.01e1.86), although these associations were attenuated. These associations were abrogated when the ISI index of insulin sensitivity was additionally added to the model. Similar results were obtained for the association between glucose-tolerance status and GGT concentrations in the highest quartile. Thus, as compared with the NGT-1h-

low group, NGT-1h-high subjects exhibited a 1.5-fold increased risk to have GGT concentrations in the highest quartile of the model (OR: 1.50; 95% CI: 1.02e2.17; P Z 0.03), whereas IGT subjects had a 1.7-fold increased risk of higher GGT concentrations (OR: 1.79; 95% CI: 1.25e2.57; P Z 0.001). Following additional adjustment for BMI (or waist circumference), SBP, DBP, triglycerides, total cholesterol and HDL, an increased risk of having GGT concentrations in the highest quartile continued to be associated with NGT-1h-high status (OR: 1.49; 95%CI: 1.01e2.23; P Z 0.05). When the ISI index of insulin sensitivity was additionally added to the model, the association between GGT concentrations in the highest quartile and glucose-tolerance status was abrogated for the NGT-1hhigh group (OR: 1.26; 95% CI: 0.80e1.98; P Z 0.30). Furthermore, NGT-1h-high subjects had a 1.5-fold increased risk of having AST concentration in the highest quartile as compared with the NGT-1h-low group (OR: 1.51; 95% CI: 1.04e2.22; P Z 0.03). Similarly, an increased risk of having AST concentration in the highest quartile was observed in IGT subjects (OR: 2.07; 95% CI: 1.44e2.97; P < 0.0001). These associations were abrogated following adjustment for BMI (or waist circumference), SBP, DBP, triglycerides, total cholesterol and HDL (OR: 1.44; 95%CI: 0.96e2.16; P Z 0.08).

Discussion Subjects with NAFLD are characterised by elevated levels of liver biomarkers such as ALT, AST and GGT, and it has been suggested that NAFLD is the most common cause of chronically elevated liver enzymes. Thus, these liver biomarkers, especially ALT, are useful surrogate measures of NAFLD for large-scale studies, and such application is also supported by data that show a significant correlation between liver biomarkers and direct measures of liver fat content [23]. A number of studies have reported that

1h-OGTT and liver enzymes elevated levels of liver enzymes, predominantly ALT and GGT, predict incident type 2 diabetes [8e11]. In addition, there is evidence that individuals with IGT not only are at increased risk of developing type 2 diabetes [1,2] but also should be considered as being at increased risk of developing NAFLD [24]. A continuous increase in liver fat content, assessed by magnetic resonance tomography, has been reported when glucose tolerance progressively deteriorates from NGT to isolated IFG, isolated IGT and IFG þ IGT [25]. These findings suggest that IGT, type 2 diabetes and NAFLD might share common pathophysiological mechanisms. Conversely, longitudinal studies have reported that a significant proportion of individuals with NGT at baseline are at risk for developing type 2 diabetes [1,2], indicating that the future risk for this clinical outcome is not similar among all subjects with NGT. In this regard, two recent studies have demonstrated that a cutoff point of 155 mg dl1 for the 1-h post-load PG during the OGTT can identify non-diabetic individuals (NGT-1h-high) who are at risk of developing type 2 diabetes [2]. With all these observations in perspective, we aimed to determine whether NGT-1h-high subjects are also characterised by elevated levels of liver biomarkers. In the present study, significant cross-sectional associations of biomarkers of NAFLD, including ALT, GGT and, to a lesser extent, AST, with elevated 1-h post-load PG during the OGTT in glucosetolerant subjects were documented. NGT-1h-high subjects are at an increased risk of having elevated concentrations of liver biomarkers of NAFLD. These associations were not modulated by gender and age, and remained unchanged after adjustment for several potential confounders including components of the metabolic syndrome such as BMI, waist circumference, blood pressure and lipid levels. However, NGT-1h-high subjects had a metabolic risk profile, including biomarkers of NAFLD, which was intermediate between that observed in NGT-1h-low subjects and that of IGT individuals, according to the relatively higher risk to develop diabetes of the latter group of pre-diabetic subjects [2]. Cut-off points for elevated ALT, AST and GGT have not been unequivocally defined, and elevation in transaminase concentrations was found in only 30% of fatty liver cases [26]. An updated definition of healthy ranges for ALT in men and women has been suggested and this has lowered the upper limits of the normal range [22]. Using these cut-off points instead of the highest quartile for the definition of elevated ALT levels, it was identified that NGT-1h-high subjects continued to be associated with an increased risk of having elevated ALT concentrations even after adjusting for BMI (or waist circumference), blood pressure and lipid levels. However, it is important to note that this lowered upper limit for ALT has been shown to be associated with a greater sensitivity, but considerably lower specificity and clinical utility, for the diagnosis of NAFLD [27]. The pathophysiological mechanisms explaining the associations between elevated liver biomarkers and 1-h post-load hyperglycaemia are unknown. Elevations of liver enzymes such as ALT and GGT or liver-secreted protein fetuin-A, a natural inhibitor of the insulin receptor tyrosine kinase in the liver and skeletal muscle, have been associated with insulin resistance [7,26,28]. In the present study, it has been identified that NGT-1h-high subjects exhibit

717 a reduction of the Matsuda index, a surrogate index of insulin sensitivity, as compared with NGT-1h-low subjects, according to previous reports [3,4]. In addition, elevated liver biomarkers might be a reflection of chronic inflammation. It has become apparent that NAFLD is a state of chronic inflammation, and a series of studies have demonstrated an association between liver biomarkers and inflammatory molecules [9,29]. Hepatic fat accumulation might induce a sub-acute inflammatory response in the liver that is similar to the adipose tissue inflammation following adipocyte lipid accumulation, and fatty change is a typical response of the liver to tumour necrosis factor-a (TNF-a) e the prototypical pro-inflammatory cytokine [30]. Thus, elevated liver biomarkers might be a reflection of local and systemic inflammation associated with low levels of anti-inflammatory hormones, such as insulin-like growth-factor 1 (IGF-1) [20]. Indeed, the authors of the present article have reported that NGT-1h-high subjects are characterised by lower levels of circulating IGF1 and higher levels of high-sensitivity C-reactive protein, as compared with NGT-1h-low subjects [3], thus suggesting that an unpaired balance between pro-inflammatory cytokines and anti-inflammatory hormones may play a pathophysiological role in the up-regulation of liver biomarkers. The results presented show that a link between post-load hyperglycaemia and biomarkers of NAFLD may have clinical implications. The importance of identifying the risk of an NGT individual to present with diabetes is related to the possibility of an early diagnosis and to the potential avoidance of type 2 diabetes-related complications, including NAFLD. Lifestyle changes in subjects at high risk for developing type 2 diabetes, such as IGT individuals, have been demonstrated to be effective in reducing the incidence of type 2 diabetes [12,13]. It is notable that these treatments are also capable of improving NAFLD and reducing liver fat content [14e17]. The findings that NGT-1h-high subjects are at increased risk for both type 2 diabetes [2] and NAFLD, as assessed by elevated levels of liver biomarkers, suggest that a cut-off point of 155 mg dl1 for the 1-h post-load PG during the OGTT may be useful to identify a subgroup of NGT individuals who may potentially harbour an increased risk of becoming diabetic; in these patients, it would be important to measure liver function, which is not routinely assessed, because they could benefit from lifestyle-change interventions and, possibly, pharmacotherapy in order to prevent or delay adverse clinical outcomes. The relatively large sample size, the inclusion of both sexes, the homogeneity of the sample with detailed metabolic characterisation and the exclusion of confounding conditions that are characterised by elevation in liver enzymes such as heavy drinking or positivity for antibodies to HCV or HBsAg are major strengths of this study. Nevertheless, the present study has certain limitations that require consideration. First, NAFLD was not assessed by invasive methods such as liver biopsy or non-invasive methods such as proton magnetic resonance spectroscopy or computed tomographic (CT) scanning. Second, because of the crosssectional design of the present study, our findings reflect only an association with prevalent, and not incident, NAFLD. Furthermore, the information on alcohol intake was assessed by a self-reported questionnaire and, thus, the actual daily alcohol consumption may have been underestimated.

718 In addition, a single 75-g OGTT was performed in participants to assess glucose levels, a common feature among most large-scale studies. These assessments are subject to intraindividual variability, and this factor may have introduced some imprecision in the classification of subjects into the glucose-tolerance groups that may have influenced the results. Finally, the present results are only based on a study in Caucasian subjects, and different findings might be observed in subjects of different ethnicity.

E. Succurro et al.

[14] [15]

[16]

Acknowledgements [17]

This work was supported, in part, by the Italian Ministry of Health (grant number RF-FSR-2007-631176) to Giorgio Sesti. [18]

References [1] Unwin N, Shaw J, Zimmet P, Alberti KGMM. Impaired glucose tolerance and impaired fasting glycemia: the current status on definition and intervention. Diabet Med 2002;19:708e23. [2] Abdul-Ghani MA, Abdul-Ghani T, Ali N, Defronzo RA. One-hour plasma glucose concentration and the metabolic syndrome identify subjects at high risk for future type 2 diabetes. Diabetes Care 2008;31:1650e5. [3] Succurro E, Marini MA, Arturi F, Grembiale A, Lugara ` M, Andreozzi F, et al. Elevated one-hour post-load plasma glucose levels identifies subjects with normal glucose tolerance but early carotid atherosclerosis. Atherosclerosis 2009; 207:245e9. doi:10.1016/j.atherosclerosis.2009.04.006. [4] Bardini G, Dicembrini I, Cresci B, Rotella CM. Inflammation markers and metabolic characteristics of subjects with 1-h plasma glucose levels. Diabetes Care 2010;33:411e3. [5] Yki-Ja ¨rvinen H. Liver fat in the pathogenesis of insulin resistance and type 2 diabetes. Dig Dis 2010;28:203e9. doi: 10.1159/000282087. [6] Tessari P, Coracina A, Cosma A, Tiengo A. Hepatic lipid metabolism and fatty liver disease. Nutr Metab Cardiovasc Dis 2009;19:291e302. doi:10.1016/j.numecd.2008.12.015. [7] Targher G. Non-alcoholic fatty liver disease, the metabolic syndrome and the risk of cardiovascular disease: the plot thickens. Diabet Med 2007;24:1e6. [8] Vozarova B, Stefan N, Lindsay RS, Saremi A, Pratley RE, Bogardus C, et al. High alanine aminotransferase is associated with decreased hepatic insulin sensitivity and predicts the development of type 2 diabetes. Diabetes 2002;51:1889e95. [9] Hanley AJ, Williams K, Festa A, Wagenknecht LE, D’Agostino Jr RB, Kempf J, et al. Elevations in markers of liver injury and risk of type 2 diabetes: the insulin resistance atherosclerosis study. Diabetes 2004;53:2623e32. [10] Fraser A, Harris R, Sattar N, Ebrahim S, Davey Smith G, Lawlor DA. Alanine aminotransferase, gamma-glutamyltransferase, and incident diabetes: the British Women’s Heart and Health Study and meta-analysis. Diabetes Care 2009;32:741e50. [11] Adams LA, Sanderson S, Lindor KD, Angulo P. The histological course of nonalcoholic fatty liver disease: a longitudinal study of 103 patients with sequential liver biopsies. J Hepatol 2005; 42:132e8. doi:10.1016/j.jhep.2004.09.012. [12] Knowler WC, Barrett-Connor E, Fowler SE, Hamman RF, Lachin JM, Walker EA, et al. Reduction in the incidence of type 2 diabetes with lifestyle intervention or metformin. N Engl J Med 2002;346:393e403. [13] Tuomilehto J, Lindstro ¨m J, Eriksson JG, Valle TT, Ha ¨ma ¨la ¨inen H, Ilanne-Parikka P, et al. Prevention of type 2

[19]

[20]

[21]

[22]

[23]

[24]

[25]

[26] [27]

[28]

[29]

[30]

diabetes mellitus by changes in lifestyle among subjects with impaired glucose tolerance. N Engl J Med 2001;344: 1343e50. Harrison SA, Day CP. Benefits of lifestyle modification in NAFLD. Gut 2007;56:1760e9. doi:10.1136/gut.2006.112094. Wang RT, Koretz RL, Yee Jr HF. Is weight reduction an effective therapy for nonalcoholic fatty liver? A systematic review. Am J Med 2003;115:554e9. Bellentani S, Dalle Grave R, Suppini A, Marchesini G. Fatty Liver Italian Network (FLIN). Behavior therapy for nonalcoholic fatty liver disease: the need for a multidisciplinary approach. Hepatology 2008;47:746e54. Kantartzis K, Thamer C, Peter A. High cardiorespiratory fitness is an independent predictor of the reduction in liver fat during a lifestyle intervention in non-alcoholic fatty liver disease. Gut 2009;58:1281e8. doi:10.1136/gut.2008.151977. Succurro E, Andreozzi F, Marini MA, Lauro R, Hribal ML, Perticone F, et al. Low plasma insulin-like growth factor levels are associated with reduced insulin sensitivity and increased insulin secretion in nondiabetic subjects. Nutr Metab Cardiovasc Dis 2009;19:713e9. doi:10.1016/j.numecd.2008.12.011. Sesti G, Succurro E, Arturi F, Andreozzi F, Laino I, Perticone M, et al. IGF-1 levels link estimated glomerular filtration rate to insulin resistance in obesity: a study in obese, but metabolically healthy, subjects and obese, insulin-resistant subjects. Nutr Metab Cardiovasc Dis; 2010;. doi:10.1016/j.numecd.2010.02.008. Succurro E, Andreozzi F, Sciaqua A, Hribal ML, Perticone F, Sesti G. Reciprocal association of plasma IGF-1 and interleukin-6 levels with cardiometabolic risk factors in nondiabetic subjects. Diabetes Care 2008;31:1886e8. Matsuda M, DeFronzo RA. Insulin sensitivity indices obtained from oral glucose tolerance testing: comparison with the euglycemic insulin clamp. Diabetes Care 1999;9:1462e70. Prati D, Taioli E, Zanella A, Della Torre E, Butelli S, Del Vecchio E, et al. Updated definitions of healthy ranger for serum alanine aminotransferase levels. Ann Intern Med 2002;137:1e9. Westerbacka J, Corner A, Tiikkainen M, Tamminen M, Vehkavaara S, Hakkinen AM, et al. Women and men have similar amounts of liver and intra-abdominal fat, despite more subcutaneous fat in women: implications for sex differences in markers of cardiovascular risk. Diabetologia 2004;47:1360e9. Yun JW, Cho YK, Park JH, Kim HJ, Park DI, Sohn CI, et al. Abnormal glucose tolerance in young male patients with nonalcoholic fatty liver disease. Liver Int 2009;29:525e9. Kantartzis K, Machann J, Schick F, Fritsche A, Ha ¨ring HU, Stefan N. The impact of liver fat vs visceral fat in determining categories of prediabetes. Diabetologia 2010;53:882e9. Stefan N, Kantartzis K, Ha ¨ring HU. Causes and metabolic consequences of fatty liver. Endocr Rev 2008;29:939e60. Kunde SS, Lazenby AJ, Clements RH, Abrams GA. Spectrum of NAFLD and diagnostic implications of the proposed new normal range for serum ALT in obese women. Hepatology 2005;42:650e6. Stefan N, Fritsche A, Weikert C, Boeing H, Joost HG, Ha ¨ring HU, et al. Plasma fetuin-A levels and the risk of type 2 diabetes. Diabetes 2008;57:2762e7. Kerner A, Avizohar O, Sella R, Bartha P, Zinder O, Markiewicz W, et al. Association between elevated liver enzymes and C-reactive protein: possible hepatic contribution to systemic inflammation in the metabolic syndrome. Arterioscler Thromb Vasc Biol 2005;25:193e7. doi: 10.1161/01.ATV.0000148324.63685.6a. Tilg H, Hotamisligil GS. Nonalcoholic fatty liver disease. Cytokine-adipokine interplay and regulation of insulin resistance. Gastroenterology 2006;131:934e45. doi:10.1053/ j.gastro.2006.05.054.