Predictive Value of Serum Gamma-glutamyltranspeptidase for Future Cardiometabolic Dysregulation in Adolescents- a 10-year longitudinal study

Serum gamma-glutamyltransferase (γ-GT) is implicated in the pathogenesis of atherosclerosis and metabolic syndrome (MetS) in adults. The relationships between γ-GT and cardiometabolic dysregulation remains unclear in adolescents. We enrolled 7,072 Taiwanese adolescents and followed them for a median of 6.8 years. The optimal cut-off values (CoVs) of baseline γ-GT to predict future MetS, hypertension (HTN), and type 2 diabetes (T2DM) were determined by receiving operating characteristic (ROC) curve. Using these CoVs, the participants were divided into normal- and high-level groups. Cox proportional hazard analysis was used to calculate hazard ratios (HRs) for the subjects with a high level of γ-GT for the risk of future cardiometabolic dysregulation. Serum γ-GT was significantly higher in the subjects with MetS than in those without MetS at baseline (p < 0.001). The optimal CoVs of γ-GT were 12 U/L for boys and 11 U/L for girls. In multivariate Cox regression analysis, a higher serum γ-GT level increased the risk of future MetS (HRs 1.98 and 2.85 for boys and girls, respectively, both p < 0.001), but not new onset HTN and T2DM. In conclusion, serum γ-GT levels not only demonstrated an excellent correlation with the presence of MetS and also in predicting future MetS in adolescents.

ratio have been reported to be useful biomarkers in establishing MetS 13 , the limited testing ability in primary care institutes limits their clinical application. With an increasing prevalence of MetS in adolescents 14 , identifying easy and reliable biomarkers to predict cardiometabolic dysregulation and understanding the relationships between these biomarkers and cardiometabolic dysregulation are also important.
Gamma-glutamyltranspeptidase (γ-GT) is a liver enzyme that participates in the synthesis and degradation of glutathione as well as xenobiotic detoxification 15,16 . Serum γ-GT is a widely used biomarker for alcoholic liver injury and nonalcoholic fatty liver disease (NAFLD). Previous studies have also reported the diagnostic role of serum γ-GT in MetS, T2DM, and CVD, and its predictive role of mortality and morbidity associated with cardiometabolic dysregulation [17][18][19][20] . However, these studies only enrolled middle-aged patients, and thus cannot be extrapolated to adolescents [17][18][19][20] . A recent cohort study recruiting 1,874 adolescents demonstrated that the subjects with NAFLD had higher γ-GT levels and greater liver shear velocity (an indicator of liver fibrosis) than those without NAFLD, even after adjustment for fat mass 21 . Although the association between serum γ-GT and ultrasound scan-determined liver damage was identified 21 , the cross-sectional study cannot determine the causality. In addition, the role of γ-GT in future cardiometabolic dysregulation is also uncertain in adolescents. This longitudinal study aimed to evaluate the relationships between baseline γ-GT levels and MetS and its component, and to assess whether optimal cut-off values (CoVs) of γ-GT can predict future MetS, hypertension (HTN) and T2DM in adolescents.

Methods
This study was approved by the Ethical Committee of the Cardinal Tien Hospital and the Ethical Committee of MJ Health Screening Centers. Each participant provided written informed consent. The described methods were carried out in accordance with the guidelines of the Declaration of Helsinki. Study Participants. We enrolled subjects from MJ Health Screening Centers, a privately-owned chain of clinics throughout Taiwan which provide regular health examinations to their members. Parental informed consent and assent form the young adolescents were obtained. Data from the participants were collected anonymously and provided for research purposes only. In total, 11,370 subjects aged from 10 to 15 years were enrolled during a 10-year sample period (1999 to 2008) (Fig. 1). The exclusion criteria were those with only one visit (n = 3,545), those with missing data of MetS components or γ-GT (n = 512) and those with a history of alcohol consumption, HTN, type 1 diabetes and those taking medications known to affect MetS components or serum γ-GT levels including antihypertensive agents, corticosteroid, glycemic control agent, antilipid agent, antipsychotics, antidepressants, antiepileptics and immunosuppressants (n = 241). The remaining 7,072 subjects (3,954 boys and 3,118 girls) were enrolled as the study cohort. Study Design. There are two parts to this study. The first was a cross-sectional observation on the relationships between baseline γ-GT levels and MetS and its components. In addition, the optimal CoVs of baseline γ-GT to differentiate the subjects with and without MetS were identified. The second stage of this study was longitudinal. The primary aim of this stage was to validate the CoVs determined in stage 1. Thus, 551 subjects who had MetS at baseline were excluded, and the remaining 6,521 subjects without MetS were followed up annually with the range of 2 to 10 years (median 6.8 years). Based on the γ-GT CoVs, we grouped the subjects without MetS into those with normal-and high-levels of γ-GT. The incidence rates of developing future MetS, HTN and T2DM were then calculated in the two groups.
General Data and Anthropometric Measurements. The senior nursing staff used a questionnaire to obtain the subjects' drinking habits and medical history. Complete physical examinations were then performed. Anthropometric measurement including waist circumference (WC), body weight, body height, systolic blood pressure, and diastolic blood pressure were measured as we described previously [22][23][24] . After 10-hour fasting, blood samples were drawn from the antecubital vein for biochemical analysis. Plasma was separated from the blood within 1 hour and stored at −30 °C until fasting plasma glucose (FPG) and lipid profile analysis. The FPG was detected using a glucose oxidase method (YSI 203 glucose analyzer, Scientific Division, Yellow Springs Instruments, Yellow Springs, OH). Total cholesterol, triglycerides (TG), and low-density lipoprotein cholesterol (LDL-C) concentrations were measured by an enzymatic colorimetric method with a Roche Cobas C501 Chemistry Analyzer (Diamond Diagnostics, USA). Serum levels of high-density lipoprotein cholesterol (HDL-C) were determined using an enzymatic colorimetric assay after dextran sulfate precipitation. Serum γ-GT levels were measured using a CX7 biochemistry analyzer (Beckman, Fullerton, CA) 22-24 . Definition of Metabolic Syndrome. We used the International Diabetes Federation (IDF) consensus definition of MetS in children and adolescents to define MetS 22,25 . Subjects having three or more of the following abnormalities were diagnosed with MetS: abdominal obesity (WC ≥ 90 th percentile) 26 , TG ≥ 150 mg/dL, HDL-C < 40 mg/dL, HTN (systolic blood pressure ≥130 or diastolic blood pressure ≥85 mmHg), and FPG concentration ≥100 mg/dL 22 . Statistical Analysis. Anthropometric and biochemical data were expressed as mean ± standard deviation.
All data were tested for normal distribution using the Kolmogorov-Smirnov test and homogeneity of variance with Levene's test. The t-test was used to evaluate differences in demographic data between the subjects with and without MetS. Univariate and multivariate regression analyses were used to assess correlations between γ-GT and MetS components. The optimal CoVs of γ-GT for a higher likelihood of developing cardiometabolic dysregulation was calculated using receiver operating characteristic (ROC) curve analysis (MedCalc Software, Broekstraat, Mariakerke, Belgium). In stage 2, hazard ratios (HRs) of having MetS, HTN and T2DM were calculated using Cox regression analysis. In addition, Kaplan-Meier plots and the log rank test were performed to evaluate the time effect on the incidence of having MetS, HTN and T2DM between the two groups. All data were analyzed using SPSS 18.0 software (SPSS Inc., Chicago, IL). A p-value (two-sided) < 0.05 was considered to be statistically significant.

Results
Baseline Characteristics and Association between γ-GT and MetS. The baseline demographic data of the participants with and without MetS are shown in Table 1. Of the 3,954 male subjects, 332 (8.4%) with a mean age of 13.31 ± 1.97 years and 219 (7.0%) of 3,118 females with a mean age of 13.47 ± 1.96 years fulfilled the diagnostic criteria of MetS. There were significant differences in all five components of MetS (WC, blood pressure, FPG, HDL-C, and TG) between the subjects with and without MetS in both genders. Notably, the level of serum γ-GT was significantly higher in the subjects with MetS than in those without (p < 0.001).
Univariate regression analysis showed a significant correlation between γ-GT and all five components of MetS in the males, however, only WC, blood pressure and TG were associated with γ-GT in the females (Table 2). In multivariate regression analysis, WC, HDL-C and TG in the males and WC and TG in the females remained significantly associated with γ-GT levels.
ROC curve analysis showed that the optimal CoVs of γ-GT were 12 U/L in males and 11 U/L in females (Fig. 2). The areas under the ROC curve were 0.68 for the males (sensitivity 74.1%, specificity 52.0%) and 0.64 for the females (sensitivity 60.3%, specificity 60.2%) (both p < 0.001). , and a history of alcohol consumption, HTN, type 1 diabetes and those taking medications known to affect MetS components or serum γ-GT levels (n = 241) were excluded. The remaining 7,072 subjects were enrolled as the study cohort. In stage 1, the optimal CoVs of baseline γ-GT to differentiate the subjects with and without MetS were identified by ROC curve. Using these CoVs, the aim of second stage was to validate its predictive role on future MetS, HTN and T2DM.  (Table 3). In addition, multivariate Cox regression analysis showed that a higher serum γ-GT level remained a significant risk factor for future MetS (HR 1.98, 95% confidence interval (CI) 1.42-2.77 in males; HR 2.85, 95% CI 1.60-5.08 in females, both p < 0.001), but not in new-onset HTN orT2DM. Kaplan-Meier plots also demonstrated the same findings (Fig. 3).

Discussion
The results of this study revealed that the adolescents with MetS not only had higher γ-GT levels, but also a significant association between γ-GT and MetS, particularly WC and TG. These findings suggest that γ-GT may be involved in the pathophysiology of MetS in adolescents. In accordance with this hypothesis, our longitudinal results over a median follow-up period of 6.8 years indicated that a high serum γ-GT level was an independent predictor for future MetS in adolescents. To the best of our knowledge, this is the first large-scale longitudinal study focusing on adolescents to investigate the role of γ-GT on future MetS, HTN and diabetes in the same time.
Since it is well-known that central obesity and insulin resistance are at the core of MetS, the role of γ-GT in the pathogenesis of MetS might be linked through NAFLD. In subjects with NAFLD, overproduction of glucose and TG from the fatty liver may precipitate the occurrence of MetS. On the other hand, the NAFLD is considered the hepatic manifestation of MetS and commonly associated with obesity 27 . Therefore, NAFLD was reported to be a useful predictor of MetS 28 . Conversely, patients with MetS have an increased risk of developing NAFLD 29 Table 2. Univariate and multivariate regression analysis of the γ-GT and components of the metabolic syndrome Model 1: Adjusted for components of metabolic syndrome. Model 2: Adjusted for components of metabolic syndrome as well as age and low-density lipoprotein.
highly increasing prevalence of T2DM, obesity, and lifestyle changes (mainly exercise withdrawal) in the general population also makes NAFLD the most common diagnosis in daily clinical practices 30 . Even though NAFLD as a cause or a consequence of MetS is still being debated, an elevated level of γ-GT secondary to excessive liver fat accumulation has been demonstrated in patients with MetS and NAFLD 17 . As expected, γ-GT has been reported to be a surrogate marker of NAFLD, and also a promising biomarker for MetS and its components in adults 19,20 . However, little is known about the associations of γ-GT concentration with MetS and the role of γ-GT as features of MetS in adolescents. To elucidate this uncertainty, Kong et al. enrolled 2,067 healthy Hong Kong participants aged 6-20 years and demonstrated that high γ-GT levels were associated with components of MetS, especially obesity and high blood pressure 31 . Even though these striking findings support the assumption that serum γ-GT might be a potential predictor for MetS in the youth population, the cross-sectional study could not provide information regarding the temporal and causal relationship between γ-GT and MetS 31 . The present study taking advantage of large-scale longitudinal follow-up aimed to assess the predictive value of γ-GT on future MetS in adolescent males and females. Interestingly, our results showed that γ-GT levels were distinctly associated with the WC, HDL-C and TG components of MetS in the males, but only WC and TG in the females. Similarly, previous studies also suggested differences in age and gender in the way MetS is expressed in adults 32 as in adolescents 33 . Even though the phenotype of MetS determined by gender was identifiable, we found that a hypertriglyceridemic waist (HTGW) was strongly related to γ-GT levels in both genders, suggesting that HTGW is a useful index for metabolic dysregulation [33][34][35] . In addition to γ-GT, our results also showed that WC in males and HDL-C in females had predictive power for new-onset MetS. These findings reinforce the hypothesis that MetS is a heterogeneous condition, so that the predictive parameters of MetS in affected subjects can be influenced by age, gender, and race/ethnicity 36 . Taken together, our compelling findings not only identify the relationships between γ-GT, current MetS and future MetS, but also validate differences in gender in the variable expression of MetS in adolescents 31,33,35 .
Emerging evidence has revealed the association between NAFLD and increasing odds of MetS 24 . The risk reduction of MetS may be achieved by lowering liver fat. Although pharmacologic therapies for NAFLD remains unavailable 37 , lifestyle interventions such as dieting and exercise have been considered effective 38 . In regard to exercise, Keating et al. reported that aerobic exercise training may help to burn fat in liver and viscera regardless of aerobic exercise dose or intensity 39 . Another study on resistance exercise also demonstrated that resistance training lead to a significant reduction in liver fat content and a greater glycemic control in the meanwhile 40 . Even   Table 3. Hazard ratios of γ-GT and components of the metabolic syndrome in developing future metabolic syndrome, hypertension and type 2 diabetes. BP, blood pressure; CI, confidence interval; FPG, fasting plasma glucose; γ-GT, gamma-glutamyltranspeptidase; HDL-C, high-density lipoprotein cholesterol; WC, waist circumference. *Criteria for WC were according to the cut-off value by Sung et al. 26 ; criteria for BP were systolic BP ≥ 130 mmHg or diastolic BP ≥ 85 mmHg; criteria for HDL-C was<40 mg/dL. BP, blood pressure; CI, confidence interval; FPG, fasting plasma glucose; γ-GT, gamma-glutamyltranspeptidase; HDL-C, high-density lipoprotein cholesterol; WC, waist circumference. *Criteria for WC were according to the cut-off value by Sung et al. 26 ; criteria for BP were systolic BP ≥ 130 mmHg or diastolic BP ≥ 85 mmHg; criteria for HDL-C was <40 mg/dL. though the existing evidences all supported the role of exercise on improving NAFLD and MetS [38][39][40] , the precise mechanisms were still unclear. On the other hand, the measurement of serum γ-GT level was a less expensive, widely available and easily interpretable way in primary care institutes to predict MetS, compared to ultrasound scan-determined NAFLD 21 . Considering the cost-effectiveness, the CoVs of γ-GT provided in the present study might be a useful tool to evaluate the long-term efficacy of exercise on NAFLD and MetS, and to clarify their relationships, at least in Taiwanese adolescents.
Although the detailed mechanisms that the link γ-GT with HTN and atherosclerotic CVD remain elusive, there are some possible explanations for their relationships 41,42 . Previous studies showed that γ-GT is significantly related to markers of inflammation such as fibrinogen, C-reactive protein and F2-isoprostanes 42,43 . Furthermore, γ-GT is thought to be involved in the pathogenesis of atherosclerosis on the basis of expression of γ-GT in human atherosclerotic lesions 43,44 . Additionally, the activity of ectoenzymatic γ-GT has been reported to play a pivotal role in the generation of free radical species through modulating the redox status of protein thiols at the cell surface 43,45 . This evidence supports the possibility that serum γ-GT is not only a marker of inflammation and oxidative stress but also a potential predictor for future HTN [42][43][44][45] .
However, the results of previous studies have been inconsistent with regards to the relationship between γ-GT and HTN. Kim et al. found a meaningful relationship between high γ-GT levels and HTN only in drinkers 46 , but Stranges et al. reported that a higher γ-GT level increased the risk of HTN in both subjects who did and did not drink alcohol 43,47 . Interestingly, our results showed that serum γ-GT levels did not have a predictive power for future HTN, suggesting a possible different pathophysiology in incident HTN in adolescents. The discrepancies between previous studies on adults and our study may be because our subjects were younger, and because they had low CoVs of γ-GT and fewer deleterious lifestyle factors (such as heavy alcohol consumption, cigarette smoking, and physical inactivity) 46,47 . Further studies including participants with a wide range of age, different genetic background, insulin resistance status, and inflammatory and oxidative condition are needed to elucidate the true role of γ-GT in predicting HTN.
Even within a normal range of concentration, serum γ-GT has been reported to be related to the presence of diabetes 17,42,48 . However, our results did not support serum γ-GT activity as a predictor of T2DM in adolescents. Several possible explanations are as follows: First, epidemiological study on prevalence of diabetes in Taiwan reported that adolescents have less than a 1% prevalence of T2DM 49 . Second, the natural time-course of diabetes is a critical confounding factor while assessing the relationship between metabolic predictors and the development of T2DM. Our subjects were relatively young so that normal glucose levels might be observed at a much earlier age in consideration of 'compensated period' , i.e., higher secretion of plasma insulin to maintain glucose homeostasis 22 . In support of this, Kong et al. have shown high γ-GT levels did not pose a significant risk Kaplan-Meier curves estimate with log rank test was applied for the event-free probability between the subjects with normal γ-GT levels (≤12 U/L) and high γ-GT levels (>12 U/L).
to dysglycaemia because of their young participants 31 . Finally, our participants were around the age of puberty, and higher levels of sex hormones may have inhibited lipogenesis and improved insulin sensitivity 50 . However, plasma insulin levels parallel to fasting glucose levels were unavailable in this study. Thus we could not evaluate the association between γ-GT and insulin resistance.
The strengths of this study include its longitudinal population-based design and the large number of participants. In addition, this is the first clinical study to identify the optimal CoVs of γ-GT in predicting future MetS in adolescents. Using this simple and widely available biomarker may be helpful in initiating preventive strategies for adolescent MetS. However, there are also several limitations to this study. First, selection bias might exist due to study participants selected from a health screening center rather than from the community. However, the aim of this study was to observe relationships between factors, and thus there should be minimal effects. Second, all subjects of our study were ethnically Chinese, limiting the generalizability of the results to other ethnicities. Finally, data on the levels of serum alanine aminotransferase, insulin, fibrinogen, C-reactive protein, adiponectin and F2-isoprostanes were lacking. Further studies including these parameters and assessing the relationships between γ-GT, systemic oxidative stress, and inflammatory status are needed.
In conclusion, the treatment and prevention of MetS in adolescents has become a public health priority. Our findings suggest that serum γ-GT levels could serve as a clinical predictor for future MetS in adolescents. Using such a low-cost and widely used metabolic biomarker may help pediatricians to screen adolescents at high risk of MetS at an early stage and prevent subsequent deleterious consequences.