A Comparative Analysis of Nutritional Assessment Using Global Leadership Initiative on Malnutrition Versus Subjective...

Malnutrition is a prevalent condition in maintenance hemodialysis (MHD) patients. This study aimed to evaluate the performance of the recently developed GLIM (Global Leadership Initiative on Malnutrition) in MHD by assessing the agreement, accuracy, sensitivity, specificity, and survival prediction of GLIM when compared to 7-point subjective global assessment (7p-SGA) and malnutrition inflammation score (MIS).

We investigated 2 cohorts: MHD_ltaly (121 adults from Italy; 67 ± 16 years, 65% men, body mass index 25 ± 5 kg/m²) and MHD_Brazil (169 elderly [age > 60 years] from Brazil; 71 ± 7 years, 66% men, body mass index 25 ± 4 kg/m²), followed for all-cause mortality for median 40 and 17 months, respectively. We applied the 2-step approach from GLIM: (1) screening and (2) confirming malnutrition by phenotypic and etiologic criteria. For 7p-SGA and MIS, a score ≤5 and ≥8, respectively, defined malnutrition.

Malnutrition was present in 38.8% by GLIM, 25.6% by 7p-SGA, and 29.7% by MIS in the MHD_Italy cohort, and in 47.9% by GLIM, 59.8% by 7p-SGA, and 49.7% by MIS in the MHD_Brazil cohort. Cohen's kappa coefficient (κ) showed only "fair" agreement between GLIM and SGA (MHD_Italy: κ = 0.26, P = .003; MHD_Brazil: κ = 0.22, P = .003) and between GLIM and MIS (MHD_Italy: κ = 0.33, P < .001; MHD_Brazil: κ = 0.25, P = .001). Cox regression analysis showed that all 3 methods were able to predict mortality in crude analysis; however in the adjusted model, the association seemed more consistent and stronger in magnitude for 7p-SGA and MIS.

In MHD patients, GLIM showed low agreement, sensitivity, and accuracy in identifying malnourished subjects by either 7p-SGA or MIS. Considering the specific wasting characteristics that predominate in MHD, the well-established 7p-SGA and MIS methods may be more useful in this clinical setting.