Development and internal validation of a risk prediction model for refractory hypoalbuminemia in peritoneal dialysis patients to guide nutritional intervention

Background Hypoalbuminaemia is closely related to poor prognosis in patients with PD. In clinical practice, we found that hypoalbuminemia is difficult to correct and persists longer in some PD patients. The primary purpose of this study was to analyze the cause of refractory hypoalbuminemia in PD patients and construct a clinical prediction model to help identify these patients at an early stage. Methods we retrospectively enrolled PD patients treated at the Second Hospital of Hebei Medical University between January 2021 and January 2023. Clinical data, laboratory results, peritoneal equilibration test (4hD/Pcr), and peritoneal protein excretion rate were collected. Patients were divided into modeling and internal validation cohorts. Risk factors were screened by LASSO regression and included in a multivariable logistic model to construct a nomogram. Model performance was evaluated by discrimination, calibration, and decision curve analysis (DCA). Results a total of 368 PD patients were included, of whom 132 (35.9%) had refractory hypoalbuminemia. Independent risk factors included diabetes mellitus, elevated C-reactive protein (CRP), and high 4hD/Pcr; protective factors included higher BMI, triglycerides ≥1.265 mmol/L, HDL-C, and LDL-C. Peritoneal protein excretion rate was significantly higher in patients with refractory hypoalbuminemia and those with diabetes, and positively correlated with 4hD/Pcr ( r = 0.6, P < 0.001). The prediction model showed excellent discrimination in the modeling cohort (AUC = 0.939) and internal validation cohort (AUC = 0.930), with good calibration and clinical net benefit. This study identifies key risk factors for refractory hypoalbuminemia in PD patients and presents a validated prediction model with high accuracy and clinical utility. Conclusion this study identifies key risk factors for refractory hypoalbuminemia in PD patients and presents a internal validated prediction model with high accuracy and clinical utility, which can guide early intensive nutritional intervention.