Nutritional indicators' performance in malnutrition diagnosis of hospitalized elderly patients.

Background: Several nutritional diagnosis methods and their relationship with clinical outcomes have been described. This study investigated malnutrition in hospitalized elderly patients (HEP) using different nutritional indicators and determined criteria to identify malnutrition and explore the variables that discriminate the risk of malnutrition.

Method: Cross-sectional study with 500 HEP; different methods of nutritional diagnosis, their relationship with clinical outcomes and criteria for defining malnutrition were investigated. The GLIM criteria for the diagnosis of malnutrition was applied in this study. In the statistical analysis, the Chi-square test, Fisher's exact test, Mann-Whitney test, univariate and multiple logistic regression and the ROC curve were used.

Results: Patients aged 65-79 years, at nutritional risk or with malnutrition, had longer hospital stays (p = 0.0099; OR = 1.047; 95% CI = 1.011; 1.084) and lower body mass index (BMI) (p < 0.0001; OR = 0.867 (1153)); 95% CI = 0.813; 0.924 (1085; 1225). Patients aged ≥80 years had a lower BMI (p = 0.0053; OR = 0.779 (1284); 95% CI = 0.653; 0.928 (1078; 1531)). Accuracy was significant in both age groups for BMI (p < 0.0001; 65-79 years and p = 0.001; ≥80 years); for the lymphocyte count (p = 0.0167; 65-79 years and p = 0.0028; ≥80 years), and for the calf circumference (CC) (p < 0.0001; 65-79 years and p = 0.001; ≥80 years). Using the GLIM criteria, 27.78% of patients were considered malnourished. CC showed good accuracy, good specificity, but low sensitivity while BMI was more accurate to detect malnutrition in both age groups.

Conclusion: CC showed good accuracy, good specificity, but low sensitivity to detect malnutrition. BMI was more accurate in both age groups to detect malnutrition.