Non-alcoholic Fatty Liver Disease And Diabetes Mellitus

Part 1: Ultrasonography-diagnosed Non-alcoholic Fatty Liver Disease and the Risk of Type 2 Diabetes Mellitus: A Cohort StudyThe majority of the longitudinal studies have concluded that non-alcoholic fatty liver disease(NAFLD) is an independently risk factor for developing type 2 diabetes(T2DM). However, most previous studies were hospital-based which included subjects undergoing health check-ups and the results could not be generalized to the general population. More importantly, most previous studies diagnosed T2 DM without 2h postprandial glucose, which might underestimate the risk of T2 DM associated with NAFLD to some extent. Meanwhile, the association between NAFLD and the risk of pre-diabetes was also speculated, but few studies have reported.Aims:To investigate whether NAFLD diagnosed by ultrasonography could predict the risk of future T2 DM and pre-diabetes in China.Methods:The cohot study with a follow-up of 5 years was based on 2007-2008 China National Diabetes and Metabolic Disorders Survey(CDMDS), which was a nationwide population-based cross-sectional survey conducted from June 2007 to May 2008. In brief, a multi-stage stratified sampling method was used to select a nationally representative sample of Chinese adults with age above 20 years. Our present cohort sample only included 1915 subjects of CDMDS in Xi’an, Shaanxi province, Northwestern China. They were invited to participate in a follow-up evaluation between June 2012 and October 2013. Of the 1915 subjects, 806(42.1%) agreed and completed the follow-up examination. We excluded subjects with any of the following reasons: excessive alcohol consumption(n = 39), missing data(n = 10) or other reasons(n = 2). We also ruled out subjects with diabetes(n = 75) or pre-diabetes(n = 172) diagnosed by oral glucose tolerance test(OGTT) and a history of antidiabetic medications at baseline. Finally, a total of 508 subjects were included as our study sample for analysis. A standard questionnaire was designed and used to collect information. Physical and lab examination(e.g. oral glucose tolerance test) was conducted. Abdominal ultrasonography was used to diagnose fatty liver. The 1999 World Health Organization diagnostic criteria were referred for diagnosing diabetes and pre-diabetes. Cox proportional hazard regression was utilized to examine the association between NAFLD and the development of diabetes and pre-diabetes.Results:The study sample included 508 non-diabetic individuals, with 42.3% males, mean age of 45.6 years and mean BMI of 23.7 kg/m2. Of those individuals, 97(19.1%) were diagnosed as NAFLD and 411(80.9%) were as non-NAFLD. Of the 508 non-diabetic individuals at baseline, 20(3.9%) developed diabetes and 85(16.7%) developed pre-diabetes. The incidence of diabetes and pre-diabetes in the NAFLD group was 10.3% and 25.8%, respectively, whereas that in non-NAFLD group was 2.45% and 14.6%, respectively. Cox proportional hazard regression showed that the multivariable-adjusted RR of diabetes and pre-diabetes in the NAFLD group was 4.462(95% CI: 1.855-10.734, P < 0.001) and 1.642(95% CI: 0.965-2.793, P = 0.067), respectively, compared with non-NAFLD group.Conclusion:We demonstrated that NAFLD was a significant predictor for future type 2 diabetes, but not for pre-diabetes, in Xi’an, China. Patients with NAFLD should be explained at the increased risk of developing diabetes in clinic and be screened regularly by oral glucose tolerance tests.Part 2: Non-alcoholic Fatty Liver Disease, Cytokeratin-18 Fragment and the Risk of Type 2 Diabetes Mellitus: A Cohort StudyAlthough several longitudinal studies have reported that NAFLD is an independently risk factor for developing type 2 diabetes mellitus(T2DM), there lacked cohort studies focusing on the histologic severity of NAFLD with the risk of T2 DM. Cytokeratin-18(CK-18), the major intermediate filament protein in the liver, could indicate hepatocyte apoptosis in NASH, and was closely related with NASH inflammation and fibrosis stage.Aims:To evaluate whether the baseline CK-18 level could predict the risk of future T2 DM in Chinese population with ultrasonography-diagnosed NAFLD.Methods:This part was the subsequent analysis of part 1. Of 806 sujects who agreed and completed the follow-up examination, 457 were finally included as study population after excluding subjects with any of the following reasons: subjects with diabetes(n = 75) or pre-diabetes(n = 172) at baseline, excessive alcohol consumption(n = 39), missing data of CK-18(n = 35) and others(n = 10), outliers of CK-18(n = 16), or other reasons(n = 2). A standard questionnaire was used to collect information. Physical and lab examination(e.g. oral glucose tolerance test) was conducted. Abdominal ultrasonography was used to diagnose fatty liver. The 1999 World Health Organization diagnostic criteria were referred for diagnosing diabetes and pre-diabetes. The M30-Apoptosense ELISA kit was used for quantitative measurement of CK-18. The NAFLD subjects were divided as NAFLD group with lower dichotomy of CK-18 and NAFLD group with higher dichotomy of CK-18. Cox proportional hazard regression was utilized to examine the association between NAFLD and the development of diabetes.Results:Of 457 individuals, 363 were non-NAFLD, 46 were NAFLD with lower dichotomy of CK-18, and 48 were NAFLD with higher dichotomy of CK-18. The median level of CK-18 at baseline was 125.3(95% CI: 57.3-476.6) U/l. The median level of CK-18 was higher in subjects of NAFLD with higher dichotomy of CK-18(235.9, 167.3-392.5) compared with that of non-NAFLD(120.6, 87.6-186.3), while the median level of CK-18 was similar in subjects of NAFLD with lower dichotomy of CK-18(105.2, 75.6-116.9) compared with that of non-NAFLD. Of the 457 non-diabetic individuals at baseline, 19 developed diabetes. The incidence of diabetes in non-NAFLD, NAFLD with lower dichotomy of CK-18 and NAFLD with higher dichotomy of CK-18 was 2.5%, 8.7% and 12.5%, respectively. Cox proportional hazard regression showed that the adjusted RR of diabetes in NAFLD with lower dichotomy of CK-18 and NAFLD with higher dichotomy of CK-18 was 3.371(95%CI:1.047-10.858, P=0.042) and 4.712(95%CI: 1.710-12.987, P=0.003), respectively, compared with non-NAFLD.Conclusion:The study indicated that higher CK-18 level was associated with higher risk of type 2 diabetes in ultrasonography-diagnosed NAFLD in China. Since both ultrasonography and CK-18 measurement are non-invasive, we suggest that physicians should firstly screen NAFLD by ultrasonography, and, if possible, then measure CK-18 to locate the higher risk of diabetes in ultrasonography-diagnosed NAFLD population for tight management.Part 3: The risk of type 2 diabetes mellitus associated with non-alcoholic fatty liver disease might be underestimated: a systematic review and meta-analysis Many previous cohort researches have indicated the presence of NAFLD carried a 2-fold risk of developing T2 DM. However, we found that 2-hour postprandial glucose(PPG) test was not used in most previous studies as an alternative way to diagnose T2 DM. Meanwhile, in our previous population-based cohort study with sufficient ascertainment of T2DM(including 2h-PPG), we identified a much higher risk(4.462, 95% CI: 1.855-10.734, P < 0.001) of T2 DM than those in most previous reports. We, therefore, hypothesized that the risk of T2 DM in patients with NAFLD might be underestimated due to the absence of 2h PPG in determining T2 DM.Aims:To perform a Meta-analysis to test whether the risk of T2 DM in patients with NAFLD is underestimated or not.Methods:We searched electronic database(Pubmed, EMBASE and ISO Web of Science) up to November 2015. Cohort studies in which the presence of NAFLD was assessed by standard objective measurements were included. Cross-sectional and case-control studies, literature reviews, letters, comments, and conference reports that were not subsequently published were excluded. Studies that used imaging to demonstrate hepatic steatosis were acceptable and included. Studies that only used elevated liver enzymes, questionnaire, or self-report to assess NAFLD were excluded. Studies that did not excluded participants with excessive alcohol drinking at baseline were excluded. Studies in which the exposure of interest was T2 DM and the outcome of interest was the incidence of T2 DM were included. For studies to be eligible for inclusion, they must have reported(or provide sufficient data to enable the calculation of) a risk estimate for T2 DM related to NAFLD, together with a 95% CI. The methodological quality of the included studies was evaluated based on the Newcastle-Ottawa Scale(NOS) for cohort studies. A random effects model(if having substantial heterogeneity) or fixed effects model(if not having substantial heterogeneity) was used to assess study heterogeneity. The Begg’s test and Egger’s test were used to evaluate publication bias, which was further assessed by using funnel plots.Results:Overall, among 21645 references, 21 articles which identified 15 studies were finally included in the review. Of them, 3 studies had sufficient ascertainment of diabetes(FBG ≥ 7.0 mmol/l or 2h PPG ≥ 11.1 mmol/l base on OGTT, or on a history of antidiabetic medications) and 12 studies did not(FBG ≥ 7.0 mmol/l, or Hb A1 c ≥ 6.5%, and/or consumption of hypoglycemic medications). In total, 114017 patients with NAFLD and 43474 subjects with no NAFLD were meta-analyzed for the association of NAFLD with risk of T2 DM. The mean NOS score for the included studies was 8.1. The pooled RR in studies without sufficient ascertainment of diabetes was 1.70(95% CI: 1.58-1.83), which was significantly lower than that in studies with sufficient ascertainment of diabetes(4.09, 95% CI: 3.19-5.25). Substantial heterogeneity was observed in the former group(I2 = 50.1%, P = 0.024) and moderate heterogeneity was observed in the latter group(I2 = 34.7%, P = 0.216). No evidence of publication bias was detected by using Begg’s test, Egger’s test and funnel plots.Conclusion:Our current meta-analysis indicated that the risk of T2 DM with NAFLD in our previous knowledge might be underestimated due to the absence of 2h PPG test in diagnosis of diabetes. Our knowledge should be updated, and much more emphasis should be put on the screening of T2 DM in patients with NAFLD, as well as the rigorous management of NAFLD to reduce the risk of future T2 DM.