Abstract: Objective To explore the method of establishing the absolute risk prediction model for elderly patients with non-alcoholic fatty liver disease (NAFLD) and osteoporosis (OP). Methods In this study, 240 patients with NAFLD admitted to our hospital from January 2020 to January 2023 were selected as the research objects, and 120 healthy people were selected as the control group. According to whether the patients were combined with OP, they were divided into the NAFLD group and the NAFLD combined OP group. Baseline data were collected. The absolute risk of nonalcoholic fatty liver disease with osteoporosis was calculated using decision tree and Gail absolute risk estimation, and the receiver operating characteristics (ROC) curve was generated using R software to identify the best cutoff value. Methods There were statistically significant differences in TG, PINP, β-CTX, BMD, GHb, OCN, and ALP among the three groups (P<0.05). Compared with the control group, the levels of PINP, β-CTX, BMD, and OCN in NAFLD group and NAFLD combined OP group were significantly reduced. The levels of TG, GHb, and ALP were significantly increased. Four variables, PINP, β-CTX, BMD, and OCN, were selected by the decision tree model as risk factors for NAFLD and OP. The 5-year risk and IQR of the control group were 2.4% and 0.132%, 23.1% and 0.255% in the NAFLD group, and 42.3%and 0.451% in the NAFLD combined OP group, respectively. The optimal cut-off value was 0.100%. The empirical set verifies that the AUC of this model is 0.826, the sensitivity is 81.25%, the specificity is 75.38%, the accuracy is 72.94%, and the accuracy is 78.36%. Conclusion This study established a prediction model of bone metabolism characteristics and absolute risk of osteoporosis in elderly patients with nonalcoholic fatty liver disease, enabling initial predictions of bone metabolism and osteoporosis risk, thereby providing guidance for disease treatment and intervention.

Key words: Nonalcoholic fatty liver, Osteoporosis, Model construction, Verify