P528 - GLOBAL LEADERSHIP INITIATIVE ON MALNUTRITION (GLIM) CRITERIA: A SENSITIVE TOOL FOR DETECTING MALNUTRITION IN CRITICALLY ILL PATIENTS

P528

GLOBAL LEADERSHIP INITIATIVE ON MALNUTRITION (GLIM) CRITERIA: A SENSITIVE TOOL FOR DETECTING MALNUTRITION IN CRITICALLY ILL PATIENTS

B. Samant^1,*, N. Sanwalka^2,3,4, V. Kothari⁵, F. A. Sampat¹, S. Singh⁵, D. Shah¹, A. Agarwala¹

¹Nutrition Therapy Department, Kokilaben Dhirubhai Ambani Hospital & Research Center, ²Department of Nutrition Research & Biostatistics, NutriCanvas, Mumbai, ³Hirabai Cowasji Jehangir Medical Research Institute, Jehangir Hospital, ⁴School of Health Science, Savitribai Phule Pune University, Pune, ⁵Critical Care, Kokilaben Dhirubhai Ambani Hospital & Research Center, Mumbai, India

Rationale: Early and timely detection of malnutrition is imperative for time intervention and patient recovery. Hence, this study aimed to assess the prevalence of malnutrition using Global Leadership Initiative on Malnutrition (GLIM) and Patient–Generated Subjective Global Assessment (PG–SGA) and determine the diagnostic accuracy of GLIM comparison to PG-SGA.

Methods: A cross-sectional study was conducted in 1242 critically ill patients (699 males) with mean age 58.1±16.5 years at a tertiary-care hospital in Mumbai, India. Patients were classified as having no malnutrition, mild-to-moderate malnutrition and severe malnutrition using the GLIM and PG-SGA criteria. Sensitivity, specificity, positive predicted value, negative predicted value, and accuracy were calculated between the 2 scales. 

Results: : Using the GLIM criteria, 267 (21.5%) were normal, 753 (60.6%) had mild to moderate malnutrition and 222 (17.9%) had severe malnutrition. Using PG-SGA, 718 (57.8%) were normal, 473 (38.1%) had mild to moderate malnutrition and 51 (4.1%) had severe malnutrition. Sensitivity of GLIM to identify malnutrition was 94.9% (95% CI = 95.6%-96.6%), specificity was 33.4% (95% CI =30.0%-37.0%), positive predictive value was 50.8% (95% CI = 49.6% - 52.4%), negative predictive value was 89.9% (95% CI = 85.9%-92.9%). Accuracy was 59.3% (95% CI = 56.6%-62.1%).

Conclusion: A very high prevalence of malnutrition was identified in critically ill patients using the GLIM criteria as compared to PG-SGA. GLIM has very high sensitivity to identify malnutrition in critically ill patients and hence should be used regularly to identify malnutrition in critically ill patients.

Disclosure of Interest: None declared