Influencing factors of oral frailty in Chinese maintenance hemodialysis patients: a Bayesian network analysis.

This multicenter cross-sectional study analyzed the prevalence and multidimensional determinants of oral frailty in maintenance hemodialysis (MHD) patients using the Health Ecology Model (HEM) and provided evidence for targeted clinical interventions. Patients undergoing MHD across four tertiary hospitals in China were included (n = 680). Independent influencing factors were identified via binary logistic regression, and a Bayesian network model was constructed using R 4.3.3 and GeNIe 4.0 for risk inference. The prevalence of oral frailty was 42.94%. Binary logistic regression identified 11 independent influencers. Poor oral health (OR = 5.288), polypharmacy (OR = 5.194), general frailty (OR = 4.465), depression (OR = 3.393), malnutrition (OR = 3.093), eating alone (OR = 2.310), age ≥60 years (OR = 2.221), and meat-dominant dietary pattern (OR = 2.061) were risk factors, whereas higher education (ORs: 0.139-0.453), higher monthly income (ORs: 0.154-0.524), and better social support (ORs: 0.036-0.203) were protective factors (all p < 0.05). Furthermore, the Bayesian network model demonstrated that polypharmacy, malnutrition, eating alone, and general frailty directly influenced oral frailty, whereas sociodemographic factors, dietary pattern, poor oral health, depression, and social support exerted indirect effects. These results suggest that early risk identification and multilevel interventions should be incorporated into routine dialysis care to mitigate the risk of oral frailty.