Abstract: PO2030
Performance of GLIM for Nutritional Assessment of Hemodialysis Patients: Comparison with Subjective Global Assessment (SGA) and Malnutrition-Inflammation Score (MIS)
Session Information
- Health Maintenance, Nutrition, and Metabolism: Clinical
October 22, 2020 | Location: On-Demand
Abstract Time: 10:00 AM - 12:00 PM
Category: Health Maintenance, Nutrition, and Metabolism
- 1300 Health Maintenance, Nutrition, and Metabolism
Authors
- Avesani, Carla Maria, Karolinska institutet Department of Clinical Sciences Intervention and Technology, Huddinge, Stockholm, Sweden
- Sabatino, Alice, Parma University Hospital, Parma, Italy
- Guerra, Alessandro, Università degli Studi di Genova, Genova, Italy
- Carrero, Juan Jesus, Karolinska Institutet, Stockholm, Stockholm, Sweden
- Rossi, Giovanni maria, Parma University Hospital, Parma, Italy
- Stenvinkel, Peter, Karolinska institutet Department of Clinical Sciences Intervention and Technology, Huddinge, Stockholm, Sweden
- Fiaccadori, Enrico, Parma University Hospital, Parma, Italy
- Lindholm, Bengt, Karolinska institutet Department of Clinical Sciences Intervention and Technology, Huddinge, Stockholm, Sweden
Background
There is a need for methods to identify and monitor malnutrition in maintenance hemodialysis (MHD) patients (pts). We assessed GLIM (Global Leadership Initiative on Malnutrition) and evaluated agreement and survival prediction of GLIM vs. SGA and MIS in MHD pts.
Methods
We investigated two cohorts, MHDltaly (121 adult pts from Italy; 67±16y, 65% men, BMI 25±5 kg/m2) and MHDBrazil (169 elderly (age>60 y) pts from Brazil; 71±7y, 66% men, BMI 25±4 kg/m2), followed for 40 (27; 46) and 17 (12; 31) months (median and 25th; 65th), respectively. GLIM comprises: 1. Screening and 2. Confirming malnutrition by phenotypic and etiologic criteria. For 1., presence of >1 criteria from protein energy wasting definition was used. Pts at risk were re-tested with GLIM’s phenotypic criteria: non-volitional weight loss or low BMI (<20 kg/m2 if <70y, or <22 kg/m2 if <70y) and reduced muscle mass (MAMC<90%). As dialysis is a catabolic procedure, all pts were positive for the etiologic criteria. For SGA and MIS, a score ≤5 and ≧8 was considered for malnutrition, respectively.
Results
Malnutrition was present in 38.8% by GLIM, 25.6% by SGA and 29.7% by MIS in the MHDItaly cohort, and in 47.9% by GLIM, 59.8% by SGA and 49.7% by MIS in the MHDBrazil cohort. Cohen's kappa coefficient (κ) showed only “fair” agreement between GLIM and SGA and MIS respectively (Table). Cox regression analysis adjusted for gender and age showed that in the MHDItaly cohort, only pts malnourished by MIS had higher risk for mortality (HR= 2.42; 95% CI 1.28 to 4.59; P=0.007) while in the MHDBrazil cohort, pts malnourished by GLIM (HR= 2.09; 95% CI 1.13 to 3.86; P=0.02), SGA (HR= 1.96; 95% CI 1.01 to 3.79; P=0.04) and MIS (HR= 2.24; 95% CI 1.20 to 4.16; P=0.01) had higher risk for mortality.
Conclusion
In MHD pts, GLIM showed low agreement with SGA and MIS, raising question on its validity and usefulness in renal care. Only malnutrition by MIS predicted mortality risk in MHDItaly cohort, but in the MHDBrazil cohort, malnutrition by all three methods predicted higher mortality risk.
Agreement vs. GLIM | MHDItaly | MHDBrazil |
GLIM x SGA, κ (P) | 0.26 (0.003) | 0.22 (0.003) |
GLIM x MIS, κ (P) | 0.33 (< 0.001) | 0.25 (0.001) |