The Severity And Related Risk Factors Of The Fatty Liver Disease In Subjects With Abnormal Glucose Metabolism

PART I The national survey of fatty liver disease in sujects with abnormal glucose metabolismObjective: To evaluate the current state, constituent ratio, severity of fatty liver disease and the relationship of liver fat and metabolic status in subjects with abnormal glucose metabolism. We survey the fatty liver disease in subjects with impaired glucose regulation (IGR), newly diagnosed type 2 diabetes (NT2DM) and known type 2 diabetes (KT2DM) in fifteen third-grade class-A hospitals all over the China.Methods: A cross-sectional survey was performed in fifteen third-grade class-A hospitals all over the China from 2010 to 201 l.The subjects were divided into four groups:normal control (NC), IGR、NT2DM and KT2DM and required to admit consecutively. The history information, demographic information, past history, anthropometric parameters, biochemical parameters were collected and liver ultrasound (common and quantitative) were performed in all subjects.3158 subjects were admitted including 898 subjects with normal glucose (male/female 344/554),503 subjects with IGR (male/female 240/263),521 subjects with NT2DM (male/female 301/220),1236 subjects with KT2DM (male/female 703/533).Results:1. Age averaged 50.8±15.1,55.7±13.7,51.7±12.5,57.7±12.0 years and BMI was 23.6±3.4 kg/m2,25.0±3.8 kg/m2,25.8±3.9 kg/m2,25.5±4.2 kg/m2 in NC, IGR, NT2DM, KT2DM groups (P all<0.01). Liver enzymes were tested in 3053 subjects.29.5%　subjects with abnormal glucose metabolism had abnormal elevated liver enzymes.The detect rate of abnormal elevated ALT, AST, GGT or any liver enzymes increased step by step in NC, IGR and NT2DM groups, but they decreased significantly in KT2DM group comparing with NT2DM group (P<0.01). The detect rates of any abnormal elevated enzymes were 22.8%,33.6%, 43.3% and 24.2%,of ALT were 7.1%,10.2%,21.7% and 8.6%,and of AST were 5.4%,7.9% 14.2% and 6.2%, and of GGT were,16.4%,25.9%,31.3% and 17.1% in NC, IGR, NT2DM and KT2DM groups respectively. They were composed mainly by NAFLD (84.7%), then alcohol (8.5%) and virus (4.3%).2. After the removal of impairment factors for liver,2780 subjects were analyzed, and we found the same trends of liver enzymes mentioned above in four groups. The BMI and DBP, TC, TG, FPG,2hPG and HbAlc levels were higher and age, diabetic duration and HDL-c and LDL-c levels were lower significantly in elevated liver enzymes group than in normal liver enzymes group (P all<0.01). Partial correlation showed, after the adjustment for sex, age and BMI, waist circumference, hip circumference, waist hip ratio, HbAlc, FPG,2hPG and TG correlated positively and HDL-c correlated negatively with ALT. In KT2DM group, after the adjustment of sex, and BMI, ALT was correlated positively with diabetic duration. AST/ALT (a simple index of liver fibrosis) was correlated positively with diabetic duration and age. Logistic regression analysis showed that TG, HbAlc were positively and age were negatively correlated with elevated liver enzymes in IGR group. BMI and TC were positively correlated with elevated liver enzymes in NT2DM group. HbAlc, TG, FPG, UA were positively and diabetic duration were negatively correlated with elevated liver enzymes in KT2DM group.3. Liver ultrasound was performed in 2837 subjects. The detect rate of fatty liver in subjects with abnormal glucose metabolism was 57.3%. It increased step by step in NC, IGR and NT2DM groups (P all<0.01), but decreased significantly in KT2DM group comparing with NT2DM group (P<0.01).They were 25.0%, 47.1%,70.1% and 56.2% in NC, IGR, NT2DM and KT2DM groups. It was composed mainly by NAFLD (86.3%), then alcohol (7.6%) and virus (3.2%).4. After the removal of impairment factors for liver, the detect rates of NAFLD by ultrasound were 25.4%,46.7%,70.2% and 55.4% in NC, IGR, NT2DM and KT2DM groups. Metabolic parameters and liver enzymes such as BMI, SBP, DBP, TC, TG, LDL-c, FPG,2hPG, HbAlc, urine ACR, ALT, AST and GGT were higher and age and HDL-c were lower significantly in NAFLD than in non-NAFLD group (P all<0.01).6. Quantitative liver ultrasound was performed in 2124 subjects. The liver fat content (LFC) was 14.3 (7.4-23.8)% in subjects with abnormal glucose metabolism and increased step by step in NC [9.8 (6.1-15.5)%], IGR [11.6 (6.8-21.5)%]and NT2DM [18.5 (9.0-28.4)%] groups(Pall<0.01), but it decreased significantly in KT2DM [13.5 (7.1-21.6)%] group comparing with NT2DM group (P<0.01). After the removal of impairment factors for liver, LFC was 9.7(6.1-15.5)%,11.7 (6.9-21.6)%、18.3 (8.9-27.5)% and 13.5 (6.8-22.5)% in NC, IGR, NT2DM and KT2DM groups.We divided the subjects by LFC quartile (Q), and found that BMI significantly increased step by step from Q2 (P<0.05).TC (P<0.05),TG, HbAlc, FPG and 2hPG (P all<0.01) increased and HDL-c (P<0.01) decreased significantly from Q3 (LFC>12.31%), and the value of glucose in Q3 met the diagnostic criteria for diabetes. They were deteriorated in Q4 (LFC>21.14%). Partial analysis showed, after the adjustment for sex, age and BMI, waist circumference, waist hip ratio, TC, TG, FPG,2hPG and HbAlc correlated positively and HDL-c correlated negatively with LFC. LFC was correlated positively with diabetic duration for KT2DM patients.Multiple linear stepwise regression analysis showed that LFC, TG, LDL-c, etc. were positively and HDL-c and female were negatively correlated with FPG. LFC, TG, age, etc. were positively and HDL-c and female were negatively correlated with 2hPG. FPG, BMI and TC were positively and female, age, and HDL-c were negatively correlated with LFC in NC group. HbAlc and BMI were positively and age and HDL-c were negatively correlated with LFC in IGR group. BMI, SBP and female were positively and age were negatively correlated with LFC in NT2DM group. 2hPG, BMI and TG were positively and age were negatively correlated with LFC in KT2DM group.Conclusion:1. The detect rates of abnormal elevated enzymes and fatty liver by ultrasound and LFC were 29.5%,7.3%, and 14.3 (7.4-23.8)% in subjects with abnormal glucose metabolism.Those of abnormal elevated enzymes were 22.8%,33.6%,43.3% and 24.2%, and of fatty liver by ultrasound were 25.0%,47.1%,70.1% and 56.2%, and LFC averaged 9.8 (6.1-15.5)%,11.6 (6.8-21.5)%,18.5 (9.0-28.4)% and 13.5 (7.1-21.6)% in NC, IGR, NT2DM and KT2DM groups respectively.2. The high detect rate of abnormal liver enzymes implied the high prevelance of NASH in subjects with abnormal glucose metabolism.3. Glucose abnormalities and NAFLD would deteriorate each other. High LFC was independent risk factor for FPG and 2hPG and closely correlated with glucose and lipid metabolic disorder. The threshold of LFC for initiating metabolic abnormalities was> 12.31%. Glucose was independent risk factor for LFC.PART II The histological features and related impact factors in subjects with abnormal glucose metabolismObjective:To investigation of severity and related risk factors of the non-alcoholic fatty liver disease in subjects with abnormal glucose metabolism.Methods:A cross-sectional survey was performed in hospitalized patients with abnormal glucose metabolism and NAFLD who were recruited from the department of endocrinology, Zhongshan hospital Fudan University, from January 2012 to January 2014. The history information, demographic information, past history, anthropometric parameters, biochemical parameters and osteocalcin were collected and the liver biopsy was performed in all subjects to know whether they had non-alcoholic steatohepatitis (NASH) or not and the severity of the inflammation and fibrosis and related risk factors.The histological features were evaluated according to NAFLD activity score (NAS,0-8) and fibrosis grading (S,0-4).Results:1.61 subjects with abnormal glucose metabolism and NAFLD were admitted to the study including 20 subjects with IGR and 41 subjects with T2DM. Age averaged 43.5±14.5years, and diabetic duration 1 (0-3) year, BMI 24.3±4.4 kg/m2, waist circumference 96.4±10.8cm, FPG 5.8±1.4mmol/L,2hPG 12.5±4.5mmol/L, HbA1c 7.1±2.0%, ALT 89(66-139)U/L, AST 58 (41-76) U/L and GGT 62 (43-90) U/L.2. Liver biopsy results show:The detect rate of NASH(NAS≧5) was 72.1%, of borderline NASH (NAS 3-4) was 21.3% and 6.6% subjects was excluded by NASH. NAS was 5.0 (4.0-6.0).45.9% of subjects had advanced fibrosis (S≧ 2),32.8% had mild fibrosis (S1),21.3% was without fibrosis (SO), S grading was 1.0 (1.0-3.0) in all subjects. The detect rate of NASH was 80.5%, of borderline NASH was 17.1%, and 55.0% of subjects had advanced fibrosis in T2DM. The detect rate of NASH was 55.0%, of borderline NASH was 30% and 25.0% of subjects had advanced fibrosis in IGR. The detect rate of NASH and fibrosis, average score of NAS and S grading were higher significantly in T2DM than in IGR(P all<0.05).3. Of 61 subjects,22 subjects were with normal liver enzymes. Liver biopsy results show:The detect rate of NASH was 57.5%, of borderline NASH (NAS 3-4) was 24.3%, and 27.3% of subjects had advanced fibrosis in all subjects. The detect rate of NASH was 64.3%, of borderline NASH was 28.6% and 35.7% of subjects had advanced fibrosis in T2DM. The detect rate of NASH was 37.5%, of borderline NASH was 25%, and 12.5% of subjects had advanced fibrosis and 25.0% had mild fibrosis (S1) in IGR.4. FPG, TG, UAER(P all<0.05), ALT and AST (P all<0.01) were higher significantly in subjects with NAS≧ 5 than those with NAS<5.5. Age, BMI, waist circumference, HbA1c, fasting insulin (INS 0), globulin, AST (P all<0.05),2hPG, and HOMA-IR (P all<0.01) were higher significantly and albumin/globulin (P<0.01) was lower significantly in subjects with S≧2 than those with S<2.6. We further divided subjects into S0, S1-2, S3-4 three groups and found 2hPG has a trend to increase with the high level of fibrosis. △CP30/△BG30 and △I30/△ BG30 were higher in S1-2 than in S0,but lower significantly in S3-4 than in SO and S1-2 (P all<0.05). Age (P<0.05) and HOMA-IR (P<0.01) were higher significantly in S3-4 than in other groups. TG was higher significantly in Sl-2 than in other groups (P<0.05). The same trends were seen in TC, HDL-c and LDL-c but without significance.7. Partial correlation analysis showed, after adjustment for sex and age, waist circumference, glycated albumin (GA), ALT, and UAER were correlated positively with steatosis grading. BMI, GA, ALT, AST, and AUC BG were positively and △CP3o、△I30 were negatively correlated with inflammation (G) grading.2hPG, AUCBG, ALT, AST and UAER were positively and osteocalcin were negatively correlated with NAS. BMI,2hPG, AUC BG, ALT, and AST were positively and △CP30 was negatively correlated with S grading.8. Multivariate analysis showed that after forcing gender and age into the model, HbAlc (OR=1.451, P<0.05), age (OR=1.070, P<0.05) and AST (OR=1.021, P<0.05) were correlated positively with G2-3, BMI (OR=1.461, P<0.01) and age (OR=1.109, P<0.01) were correlated positively with advanced fibrosis (S≧2).Conclusion:1. The prevalence of the severe histological features (NASH and fibrosis) of NAFLD is high in subjects with abnormal glucose metabolism. The detect rate of NASH(NAS ≧ 5) was 72.1%, of borderline NASH (NAS 3-4) was 21.3%,45.9% of subjects had advanced fibrosis (S≧2),2. Normal liver enzyme could not exclude inflammation and fibrosis in the liver.3. Inflammation, fibrosis, and glucose abnormalities would deteriorate each other. High glucose was independent risk factor for liver inflammation. Liver inflammation and fibrosis increased metabolic disorders and insulin resistance. Islet beta cells secreted compensatory in the early, following by deterioration.PART III Evaluate the different values of non-invasive scores in diagnosing NASH or fibrosis in patients with abnormal glucose metabolismObjective:We used NAFLD activity score (NAS) from liver biopsy to evaluate different values of non-invasive scores in diagnosing NASH or fibrosis in patients with abnormal glucose metabolism.Methods:Subjects were the same as part Ⅱ. We used areas under the Receiver Operating Characteristic (ROC) curve to evaluate different values of non-invasive scores, and calculated sensitivity, specificity, positive predictive value (PPV) and negative predictive value (NPV). Noninvasive scores including:NAFLD Fibrosis Score, AST/ALT, NAFLD liver fat score, API score, BARD score and FIB-4 score.Results:1. The areas under the ROC curve for NAFLD liver fat score for G2-3 was 0.741 (P<0.01). The optimal cut-off was 3.037.The sensitivity was 85.7%, specificity 63.1%, PPV 48.3% and NPV 93.5%. The areas under the ROC curve for API was 0.734 (P<0.01). The optimal cut-off was 0.546. The sensitivity was 78.6%, specificity 60.9%, PPV 34.5%, and NPV 90.6%. The areas under the ROC curve for FIB4 was 0.703 (P<0.05). The optimal cut-off was 1.401. The sensitivity was 64.3%, specificity 71.7%, PPV 39.1% and NPV 86.8%. NAFLD fibrosis score, BARD score and AST/ALT had no diagnostic value for G2-3 (P all> 0.05). The NAFLD liver fat score best predicted G2-3 based on the area under the ROC curve.2. The areas under the ROC curve for NAFLD liver fat score for advanced fibrosis (S≧2) was 0.725 (P<0.01). The optimal cut-off was 4.134. The sensitivity was 46.7%, specificity 93.7%, PPV 86.7% and NPV 66.7%. The areas under the ROC curve for NAFLD fibrosis score was 0.691 (P<0.05). The optimal cut-off was-1.766. The sensitivity was 85.7%, specificity 50.0%, PPV 58.5%, and NPV 80.0%. The areas under the ROC curve for FIB4 was 0.673 (P<0.05). The optimal cut-off was 0.995. The sensitivity was 78.6%, specificity 66.2%, PPV 59.5% and NPV 75.0%. The areas under the ROC curve for API was 0.660 (P<0.05). The optimal cut-off was 0.587. The sensitivity was 64.3%, specificity 75.0%, PPV 66.7% and NPV 70.6%. BARD score and AST/ALT had no diagnostic value for G2-3 (P all> 0.05). The NAFLD liver fat score best predicted advanced fibrosis based on the area under the ROC curve.Conclusion:1. NAFLD liver fat score, API score and FIB4score could effectively diagnose the G2-3 and NAFLD liver fat score, NAFLD fibrosis score, FIB4score and API score could effectively diagnose the advanced fibrosis in the patients with abnormal glucose metabolism.2. The NAFLD liver fat score best predicted G2-3.The NPV of NAFLD liver fat score, API score and FIB4 score was high, thus the scores could exclude patients without moderate or severe inflammation.3. The NAFLD liver fat score best predicted advanced fibrosis.The second best was NAFLD fibrosis score. NAFLD liver fat score had the highest PPV, and NAFLD fibrosis score had the highest NPV. We recommended combined use of these two scores to evaluate the severity of NAFLD in the patients with abnormal glucose metabolism.