RESUMEN
OBJECTIVES: To assess the concurrent and predictive validity of different combinations of Global Leadership Initiative on Malnutrition (GLIM) criteria in patients with colorectal cancer considering different indicators of reduced muscle mass (MM) and the effects of the disease. METHODS: A secondary analysis with patients with colorectal cancer. The reduced MM was assessed by arm muscle area, arm muscle circumference, calf circumference, fat-free mass index, skeletal muscle index (SMI) and skeletal muscle. Cancer diagnosis or disease staging (TNM) was considered for the etiologic criterion referred to as the effect of the disease. The other phenotypic and etiologic criteria were also evaluated, and we analyzed 13 GLIM combinations. Concurrent validity between GLIM criteria and Patient-Generated Subjective Global Assessment was evaluated. Logistic and Cox regression were used in the predictive validation. RESULTS: For concurrent validity (n = 208), most GLIM combinations (n = 6; 54.5%) presented a moderate agreement with Patient-Generated Subjective Global Assessment and none showed satisfactory sensitivity and specificity (>80%). Reduced MM evaluated by SMI and SMI were present in the GLIM combinations associated with postoperative complications (odds ratio, ≥2.0), independent of other phenotypic and etiologic criteria. The combinations with reduced MM considering any method and fixed phenotypic criteria and TNM were associated with mortality (hazard ratio, ≥2.0). CONCLUSIONS: Satisfactory concurrent validity was not verified. The GLIM diagnosis of malnutrition was associated with postoperative complications and mortality.