Abstract: Objective To constructed risk prediction model of arteriosclerotic cardiovascular disease (ASCVD) for type 2 diabetes mellitus (T2DM) with nonalcoholic fatty liver disease (NAFLD) based on national metabolic management center (MMC). Methods The clinical data of 391 T2DM patients with NAFLD recorded in the MMC platform of Xuzhou First People′s Hospital from September 2021 to January 2024 were retrospectively analyzed, they were randomly divided into training group (274 cases) and verification group (117 cases) according to a 7∶3 ratio. They were continuously followed up on the MMC platform continued to following up from the day of discharge to August 2024. The training group was divided into ASCVD group and ASCVD group according to the presence or absence of ASCVD during the following up period, the clinical data of the 2 groups were compared, and Cox regression method was used to analyze the influencing factors of ASCVD. Regression coefficient, standard error, hazard ratio (HR) and 95% confidence interval (95%CI) were calculated. Based on the influencing factors, the risk prediction nomogram model of ASCVD in T2DM patients with NAFLD was constructed. Based on the data from the training and validation groups, the receiver operating characteristic (ROC) curve and Hosmer-Lemeshow goodness of fit analysis were used to verify the prediction efficiency and calibration degree of the model, and decision curve analysis (DCA) was used to evaluate the clinical benefit of the model. Results The following-up period was ranged from 6 to 42 months, with a median of 22 (10,35) months. The incidence of ASCVD was 31.02% (85/274) in the training group, 30.77% (36/117) in the verification group, and 30.95% (121/391) in the overall incidence. Urinary albumin to creatinine ratio (UACR) (HR=2.328, 95%CI: 1.498~3.618), atherogenic index of plasma (AIP) (HR=2.237, 95%CI: 1.505~3.323), triglyceride glucose-body mass index (TyG-BMI) (HR=2.787, 95%CI: 1.626~4.778), framingham risk score (FRS) (HR=2.497, 95%CI: 1.376~4.531), visceral-to-subcutaneous fat ratio (VSR) (HR=2.809, 95%CI: 1.564~5.048) and pulse wave velocity (PWV) (HR=3.050, 95%CI: 1.619~5.744) were the risk factors of ASCVD (P<0.05), and estimated glomerular filtration rate (eGFR) (HR=0.396, 95%CI: 0.262~0.599) was the protective factors for ASCVD (P<0.05). Based on the above 7 influencing factors, the risk prediction model of ASCVD in T2DM patients with NAFLD was constructed. The area under ROC curve for predicting ASCVD in the verification group and the training group were 0.959 (95%CI: 0.929~0.980) and 0.890 (95%CI: 0.819~0.941), the sensitivity were 87.06% and 83.33%, and the specificity were 93.65% and 92.59%. There was no significant difference between the predicted probability of ASCVD and the actual probability in the verification group and the training group (P>0.05). The model predicted that the threshold probability of ASCVD in the validation group and the training group was 0.05~0.87 and 0.04~0.85. Conclusion Based on the discovery from the MMC platform discovery, UACR, AIP, TyG-BMI, FRS, VSR and PWV are risk factors for ASCVD in T2DM patients with NAFLD, while eGFR is a protective factor, the risk prediction nomogram model based on these influencing factors has high predictive value for ASCVD.

Key words: Type 2 diabetes mellitus, Nonalcoholic fatty liver disease, Atherosclerotic cardiovascular disease, National metabolic management center