Associations Between Nonalcoholic Fatty Liver Disease And Liver Enzymes Levels With Incident Type 2 Diabetes Risk In Cohort Study

Diabetes is one of the most threateningnoncommunicable diseases that do harm to the health of human beings in the word. With the improvement of the living condition, the incidence and prevalence of diabetes is increasing. By 2015, there are 415 million diabetes in the world and the number will rise to 642 million in 2040. How to significantly reduce the incidence of diabetes is a major public health in the word.Nonalcoholic fatty liver disease(NAFLD)was defined as that there is evidence of hepatic steatosis either by imaging or by histology and there are no causes for secondary hepatic fat accumulation such as significant alcohol consumption, use of steatogenic medication or hereditary disorders. Numerous prospective studies confirmed that NAFLD is associated with the development of type 2 diabetes, independently of traditional risk factors. Meanwhile,several cross-sectional studies indicated that the prevalence of NAFLD was significantly higher in T2 DM patients. However, few prospective researches were conducted to confirm the association between type 2 diabetes and the incident risk of NAFLD. Meanwhile, few studies simultaneously investigated the bidirectional associations between NAFLD and type 2diabetes in a prospective setting. The most common characteristic of the NAFLD is that liver enzymes levels are elevated, for which the increased liver enzymes levels have been used as one of the criterias to diagnose NAFLD.The liver enzymes, namely alanine aminotransferase(ALT) and aspartate aminotransferase(AST) play a major role in maintaining fasting andpostprandial glucose concentrations.Several epidemiological studies have examined the associations between ALT levels with risk of type 2 diabetes. However, studies on the associations between AST levels and incident type 2 diabetes risk were limited and inconsistent.Numerous studies demonstrated that lifestyle interventions can effectively prevent or delay the development of diabetes in individuals at high risk. Identification of the individuals at higher risk to develop type 2 diabetes and adoption of preventive interventions was important strategy to reduce the type 2 diabetes incidence. Till now, numerous prediction models for the risk of diabetes were established to predict diabetes occurrence, such as Framingham risk score, the Finnish diabetes risk score(FINDRISC), the Indian diabetes risk score(IDRS), the German diabetes risk score, Danish diabetes risk score, and Chinese diabetes risk score. However, adding more effective risk factors to the models and improving the predictive validity of the models remained to be further investigated, especially among a middle-aged and older Chinese population.Therefore, the present study mainly explored the bidirectionalassociations between NAFLD and type 2 diabetes among a middle-aged and older Chinesepopulation. Furthermore,we examined the association between serum ALT and AST levels with the incidence of type2 diabetes and explored the prediction roles of ALT in the discrimination of type 2 diabetes.Part ? Bidirectional associations between nonalcoholic fatty liver disease and type 2diabetesObjectives:Numerous prospective studies demonstrated that NAFLD was independently and strongly associated with increased incidence type 2 diabetes. Several cross-sectional studies suggested the positive association between type 2 diabetes and the NAFLD risk. The aim of this study is to examine the bidirectional associations between NAFLD and diabetes in Chinese adults.Methods:The data was derived from the Dongfeng-Tongji cohort study, which wasestablished in 2008.We recruited27,009 retired employees who agreed to answer the questionnaire information and provided baseline blood samples.The first follow-up was conducted from April to October in 2013 and recruited 38,295 participants. Among them,25,978 participants were from the 2008 baseline data(27,009 participants) and the follow-up rate is 96.2%.NAFLD was classified as none, mild and moderate/severe based on ultrasound examinations.Participants were excluded if they had T2DM(n = 5,173), chronic hepatitis(n=783), consumption excessive alcohol(n=801) or hepatic cyst(n=1,848) at baseline.Individuals with missing data on abdominal B-type ultrasound inspection(n=804),T2DMstatus(n=120), BMI or waist circumference(n= 506) at baseline were further excluded.The analysis of NAFLD and incident diabetes risk included 16,293 participants free of diabetes at baseline.Cox proportional regression model was used to calculate the hazard ratios(HRs) for the association. Participants with missing data on abdominal B-type ultrasound inspection(n=1163) were excluded. Participants were excluded if they had fatty liver(n =8,408) or chronic hepatitis(n = 621) or hepatic cyst(n=1,880) or consumption excessive alcohol(n=557) at baseline. Individuals with missing data on abdominal B-type ultrasound inspection(n=832), T2DM(n=11), BMI or waist circumference(n=360) at baseline were further excluded. The analysis of diabetes and incident NAFLD risk included 10,759 participants free of NAFLD at baseline, and logistic regression model was used to calculate the odd ratios(OR) of NAFLD.Results: Compared with those without NAFLD, individuals with mild and moderate/severe NAFLD had a monotonic elevated risk of developing diabetes(HR:1.91[95% CI: 1.63-2.33] and 2.38[1.86-3.04], respectively) after adjustment for potential confounders. In a parallel analysis, compared to participants with fasting glucose < 6.1mmol/L, the ORs of developing NAFLD in subjects with impaired fasting glucose and diabetes were 1.33(95% CI:1.13-1.56) and 1.44(95% CI:1.24-1.68), respectively.Conclusions:Our results provide compelling evidence that the NAFLD-diabetes association is bidirectional in Chinese population.As the severity of NAFLD increased, the incidence of diabetes consistently increased. Similarly,type 2 diabetes is also associated withincreased risk of NAFLD.Part II Serum liver enzymes levels predict type 2 diabetes riskObjectives:The aim of this study is to examine the associations between serum ALT and AST levels with the incidence of type 2 diabetes amonga middle-aged and older Chinesepopulation and to explorewhether the addition of ALT and AST to the traditional risk factors could improve the predictive ability of type 2 diabetes risk.Methods: At baseline we recruited 27,009(out of 31,000) retired employees(average aged 63.6 years), Of the 27,009 subjects,25,978 participates(96.2%) finishedthe first follow-up examinationsconducted from April to October in 2013.Participantsdiagnosed with diabetes at baseline(n=5,173),consumption excessive alcohol(n=829) orincomplete data on fasting plasma glucose(n=226), ALTor AST(n=1,967)were excluded from this investigation.In addition, subjects with HBs Ag positive(n = 648) were further excluded.A total of 17,887 participants without type 2 diabetes at baseline were includedand1,159 participants developed diabetes after 4.49(0.63) years of follow-up. Cox proportional hazard regression model was used to calculate the hazard ratios(HRs) for the association between serum ALT and AST levels with diabetes.Receiver-operating characteristic(ROC) curves analysis were used to calculate the corresponding area under the ROC curve(AUC) to evaluate the predictive accuracy discrimination of models incorporating traditional risk factors with and without ALT.Results: Compared with the lowest quartile of ALT, the highest quartile had a significantly higher relative risk of type 2 diabetes(HR: 2.08, 95% CI: 1.69-2.56)after adjustment for confounders. Nosignificant associations were observed between AST and development of type 2 diabetes in the final multivariable model. Compared with the first quartile,the HRs of developing diabetesin quartiles 2~4 of AST levels were 0.96(95% CI:0.80-1.16),1.04(95% CI: 0.87-1.25) and1.18(95% CI: 0.99-1.41), respectively.The addition of ALT to the traditional risk factors did not improve the predictive ability of type 2 diabetesrisk, with AUC increasing from 0.771 to 0.773(P = 0.85).Conclusions: Our results suggest that high level of serum ALT was independently associated with highere incident type 2 diabetes risk in a middle-aged and older Chinese population. However, addition of ALT to the traditional type 2 diabetes risk prediction model did not improve the discrimination of type 2 diabetes.In conclusion, our results suggest thatthe NAFLD-diabetes association is bidirectional in a middle-aged and older Chinese population. Meanwhile, we found that high level of serum ALT was independently and postively associated with incident type 2 diabetes risk. However,addition of ALT to the traditional type 2 diabetes risk prediction model did not improve the discrimination of type 2 diabetes.