Abstract: Objective The aim of this study was to investigate the risk factors for liver fibrosis (LF) in patients with diabetes mellitus type 2 (T2DM) and non-alcoholic fatty liver diseases(NAFLD). Methods 482 patients with T2DM were recruited in our hospital between April 2020 and April 2022, and the NAFLD was diagnosed based on ultrasonography and the LF was determined by the Young's modulus. The Lasso regression model and ordinal multicategory Logistic regression were applied to reveal the risk factors for LF, and the receiver operating characteristic curves(ROC) was drawn to assess the predictive efficacy of risk factors for LF in patients with NAFLD. Results Out of the 482 patients with T2DM in our series, the concomitant NAFLD was found in 276 cases(57.3%), including LF in 58 cases(12.0%) in the latter; the age, incidence of peripheral neuropathy (PN), body mass index and Young's modulus in patients with T2DM/NAFLD/LF were (46.0±4.2)yr, 82.8%, (28.3±3.2)kg/m² and (10.8±2.2)kPa, all significantly higher than in patients with T2DM/NAFLD or in patients with T2DM; blood TG, UA, HOMA-IR, ratio of triglyceride glucose/body mass index (TyG-BMI), serum TNF-α and IL-6 level in patients with T2DM/NAFLD/LF were(2.6±0.8)mmol/L, (355.2±24.6)μmol/L, (2.6±0.4), (270.8±18.5), (8.2±2.3)pg/ml and (13.6±2.4)pg/ml, all significantly higher than in patients with T2DM/NAFLD or in patients with T2DM; the LASSO and multivariate Logistic regression analysis showed that the ratio of TyG-BMI(OR=1.012, P=0.000), IL-6(OR=2.782, P=0.000), TNF-α(OR=1.008, P=0.026), UA(OR=1.530, P=0.000) and PN (OR=1.855, P=0.010) were the independent risk factors for the severe NAFLD in patients with T2DM; the ROC analysis demonstrated that the AUCs were 0.897, 0.928, 0.840 and 0.762, when the ratio of TyG-BMI, IL-6, TNF-α and UA were applied to predict LF in patients with T2DM/NAFLD. Conclusions The LF could deteriorate the disease in patients with T2DM/NAFLD, and the clinicians should take the risk factors into consideration and the early intervention might improve the prognosis.

Key words: Non-alcoholic fatty liver diseases, Type 2 diabetes mellitus, Liver fibrosis, Lasso regression model, Diagnosis