#5405 MALNUTRITION DIAGNOSIS IN CKD PATIENTS

Abstract

Abstract Background and Aims Malnutrition is a prevalent condition in patients with chronic kidney disease (CKD) receiving replacement therapy by peritoneal dialysis (PD). Geriatric nutritional risk index (GNRI) is an effective tool for screening nutritional status in maintenance hemodialysis patients. The purpose of this study was to determine the information content of GNRI for the diagnosis of protein-energy malnutrition in patients with CKD receiving replacement therapy with PD when compared to malnutrition inflammation score (MIS) and 7-point subjective global assessment (7p-SGA). Method A prospective cohort study included 222 adults PD-patients (man 100, female 112, age 44±14). All patients received continuous ambulatory peritoneal dialysis. Median duration PD before inclusion in the study was 12 months, the observation period of patients was 36 months. Nutritional status assessment included anthropometric (body mass index, triceps skinfold thickness, midupper arm muscular area) and biochemical (albumin, C-reactive protein, total iron-binding capacity, total cholesterol, hemoglobin and others) examinations as well as bioelectrical impedance analysis (BIA – fat mass). GNRI calculated according to the formula GNRI = 1.489 × albumin (g/dL)] + [41.7 × (body wt/ideal body wt)], where the ideal body weight for women is: height (cm)–100–[(height–152) × 0.2], ideal body weight for men is equal to: height (cm) -100 - [(height - 152) × 0.4]. We used MIS and 7p-SGA as a reference standard, a score ≥6 and ≤5, respectively, defined malnutrition. The cutoff of GNRI the diagnosis of malnutrition were derived from these ROC-analysis. Results Most of the individual nutritional indexes, including the anthropometric, biochemical and BIA indexes, were significantly (P<0.001) lower in the patients with MIS score ≥6 and 7p-SGA score ≤5, thus both nutritional screening tools were considered reasonable as a reference standard to determine the information content GNRI in PD-patients. The GNRI fluctuated in the range 66–126 (median 99), MIS score – 2–24 (median 7) and 7p-SGA score – 2–7 (median 5). The GNRI showed a significantly negative correlation with MIS (r = -0,708, p<0.0001) and a significantly positive correlation with 7p-SGA (r = 0.636, p<0.0001) (Fig. 1). The most accurate GNRI cutoff to identify a malnourished patient according to the MIS was ≤99 (Fig. 2). The frequency of malnutrition among the observed patients was 52.3% when using GNRI≤99, 55.9% when using MIS≥6 and 51.4% when using 7p-SGA≤5 (n.s.). The values for sensitivity and specificity with a GNRI of ≤99 in predicting malnutrition based on the MIS were 77.4% (95%CI 69.0-84.4%) and 79.6% (95%CI 70.3-87.1%) respectively. Positive predictive value and negative predictive value also had high scores respectively 73.7% (95%CI 64.9-80.9%) and 72.2 (95%CI 63.1-79.8%). The application of the GNRI to the PD-patients found significant (p<0.001) differences in the various nutrition-related indexes (anthropometric, biochemical and BIA) between the group with a GNRI≤99 and that with GNRI>99. Mortality from all causes among patients with GNRI>99 was one case per 54.9 patient-years, among patients with GNRI≤99 was one case per 29.7 patient-years. Conclusion In PD-patients GNRI closely correlates with MIS and 7p-SGA, characterized by high sensitivity and specificity. GNRI can be used to determine the nutritional status of PD patients.