Prediction of 5-year mortality risk by malnutrition according to the GLIM format using seven pragmatic approaches to define the criterion of loss of muscle mass

Objectives: To assess the association between baseline malnutrition according to the GLIM format, using seven pragmatic approaches to define the criterion of loss of muscle mass, with mortality in the SarcoPhAge (Sarcopenia and Physical Impairment with advancing Age) study during a 5-year follow-up. Secondarily, to calculate diagnostic performance indicators, concordance, and feasibility of these 7 pragmatic approaches compared to the original GLIM criteria. Methods: Post-hoc analysis of the SarcoPhAge cohort, which included 534 community-dwelling volunteers ≥65-year-old, followed-up from 2013 to 2019. Baseline malnutrition was defined by GLIM criteria and 7 approaches: 1) Omission of a reduced muscle mass as a criterion; 2) Substitution for handgrip strength, 3) Calf-circumference, 4) Mid-arm circumference, 5) Goodman's grid, 6) Ishii's score chart, and 7) Yu's formula. The association between malnutrition (according to GLIM criteria and the 7 approaches) and mortality was assessed by Cox-regressions. Sensitivity, Specificity, Positive (PPV), Negative (NPV) predictive values, area under the curve (AUC), Cohen–kappa coefficient, and TELOS-feasibility score were calculated. Results: Data to calculate GLIM criteria were available for 373 subjects (73.07 ± 5.96 years, 56% women). Prevalence of malnutrition with GLIM criteria was 24.4% (ranged from 13.9% to 20.9% with the 7 approaches). GLIM criteria showed a HR = 3.38 (1.89–6.09) to predict mortality during the 5-year follow-up, which ranged from HR = 2.72 (1.51–4.91) to 3.94 (2.14–7.24) with the 7 approaches. All 7 approaches were feasible (TELOS ≥ 3), showed sensitivity ≥ 65%, specificity ≥ 95.4%, PPV ≥ 85%, NPV ≥ 88%, AUC ≥ 0.7 and had almost-perfect/strong concordance (k ≥ 0.7) with the original GLIM criteria. Conclusions: GLIM criteria and the 7 approaches predicted three-to four-fold mortality, all ensured an accurate diagnosis, and were feasible in clinical settings.