The Correlation Between Non-alcoholic Fatty Liver Disease Degrees Changes And Risk Of Type 2 Diabetes

AimsIt has been proved that non-alcoholic fatty liver disease(NAFLD)is significantly associated with increased risk of incident type 2 diabetes(T2DM),and the risk increases along with the development of NAFLD severity.It is unclear that the influence of changes of(NAFLD)severity on T2 DM incidence in Chinese population.Therefore,we designed this prospective study to evaluate comprehensively whether changes of NAFLD severity could make differences to the incidence of T2 DM in a large-scale adult population.MethodsA total of 48,117 individuals without T2 DM at baseline were recruited from 2010 to 2016 and their health status were examined annually during the follow-up period.T2 DM were diagnosed using standard criteria(America Diabetes Association,2014).NAFLD and its severity changes were annually detected by using standard ultrasonography criteria.The association between NAFLD severity changes and T2 DM incidence were estimated using adjusted Cox proportional hazards regression models.ResultsAmong 48,742 participants enrolled into cohort analysis,742 incident T2 DM cases during 149,278 person-years of follow-up were documented.After adjustment for potential confounding factors,hazard ratios(95% confidence intervals)(HRs [95% CIs])and P values of T2 DM incidence were 1.79(1.34,2.38)(P <0.0001),0.60(0.39,0.92)(P =0.0194),2.32(1.83,2.93)(P <0.0001)and 2.91(1.67,5.08)(P =0.0002)in subjects who newly developed mild NAFLD,who got their mild NAFLD regressed,unchanged and progressed as compared to subjects who had no NAFLD from baseline to follow-up,respectively.In male participants,After adjustment for potential confounding factors,hazard ratios(95% confidence intervals)(HRs [95% CIs])of T2 DM incidence were 1.70(1.20,2.41)(P =0.003),0.52(0.31,0.88)(P =0.0139),2.09(1.58,2.76)(P <0.0001)and 2.92(1.56,5.45)(P =0.0008)in subjects who newly developed mild NAFLD,who got their mild NAFLD regressed,unchanged and progressed as compared to subjects who had no NAFLD from baseline to follow-up,respectively.While in female participants,After adjustment for potential confounding factors,hazard ratios(95% confidence intervals)(HRs [95% CIs])of T2 DM incidence were 1.99(1.19,3.32)(P =0.0089),0.79(0.37,1.70)(P =0.5411),2.92(1.91,4.46)(P <0.0001)and 3.02(0.85,10.73)(P =0.0875)in subjects who newly developed mild NAFLD,who got their mild NAFLD regressed,unchanged and progressed as compared to subjects who had no NAFLD from baseline to follow-up,respectively.While compared to those whose mild NAFLD unchanged,HRs(95% CIs)of T2 DM incidence were 0.26(0.17,0.39)(P <0.0001)and 1.26(0.76,2.07)(P =0.3675)in subjects who got mild NAFLD regressed and progressed after adjustment for potential confounding factors,respectively.In male participants,compared to those whose mild NAFLD unchanged,HRs(95% CIs)of T2 DM incidence were 0.25(0.16,0.41)(P <0.0001)and 1.40(0.81,2.43)(P =0.2342)in subjects who got mild NAFLD regressed and progressed after adjustment for potential confounding factors,respectively.In female participants,compared to those whose mild NAFLD unchanged,HRs(95% CIs)of T2 DM incidence were 0.27(0.13,0.56)(P =0.0004) and 1.04(0.32,3.37)(P =0.9537)in subjects who got mild NAFLD regressed and progressed after adjustment for potential confounding factors,respectively.ConclusionsIndividuals with development of new mild NAFLD,mild NAFLD sustained,mild NAFLD progressed were associated with increased risk of T2 DM incidence as compared to non-NAFLD ones.In contrast,regression of existing NAFLD was associated with decreased risks of T2 DM incidence whether compared to mild NAFLD sustained ones or non-NAFLD ones.