Geriatric nutritional risk index combined with modified creatinine index for the prognostic assessment of elderly patients with maintenance hemodialysis

To explore the value of geriatric nutritional risk index (GNRI) combined with modified creatinine index for the prognostic assessment of elderly patients with maintenance hemodialysis (MHD). A total of 129 elderly MHD patients were enrolled. GNRI and modified creatinine index were calculated, and patients were followed for all-cause mortality. Risk factors for overall survival were identified using univariate and multivariate Cox regression. The predictive value of GNRI and modified creatinine index, alone or in combination, was assessed by time-dependent receiver-operating characteristic (ROC) curves. The median hemodialysis vintage was 30 months, and the median follow-up was 28 months. A total of 33 patients (25.58%) reached the primary end-point. The Kaplan-Meier survival curve showed that the overall survival rate of the GNRI > 91.2 group was significantly better than that of the GNRI < 91.2 group (HR = 3.28, P = 0.002); the overall survival rate of the modified creatinine index > 20.9 group was significantly better than that of the modified creatinine index ≤ 20.9 group (HR = 1.83, P = 0.03). Deceased patients were older, had longer dialysis vintage, higher rates of hypertension, diabetes, and stroke, and exhibited lower BMI, hemoglobin, GNRI, and higher modified creatinine index. Multivariate analysis confirmed that GNRI < 91.2 (HR = 2.01, 95%CI 1.23–3.09) and modified creatinine index ≤ 20.9 (HR = 1.71, 95% CI 1.05–3.10) were independent prognostic factors. Time-dependent ROC analysis at 24 and 30 months showed AUCs of 0.79, 0.58, and 0.80, and 0.74, 0.61, and 0.74 for GNRI, modified creatinine index, and their combination, respectively. DeLong’s test indicated that the combined index did not significantly outperform GNRI alone at any time point ( P > 0.05), but was significantly superior to the modified creatinine index alone during the 24–28 month period ( P < 0.05). Low GNRI was associated with shorter overall survival in elderly MHD patients, and was an independent prognostic factor. Adding the modified creatinine index did not significantly improve the predictive value of GNRI alone, indicating that GNRI alone is sufficient for effective risk stratification. Not applicable.