Abstract
BACKGROUND: Reliable and valid tools to screen for malnutrition in the intensive care unit (ICU) remain elusive. The sarcopenia index (SI) [(serum creatinine/serum cystatin C) × 100], could be an inexpensive, objective tool to predict malnutrition. We evaluated the SI as a screening tool for malnutrition in the ICU and compared it with the modified-NUTRIC score. MATERIALS AND METHODS: This was a historical cohort study of ICU patients with stable kidney function admitted to Mayo Clinic ICUs between 2008 and 2015. Malnutrition was defined by the Subjective Global Assessment. Diagnostic performance was evaluated with the area under the receiver operating characteristic curve (AUC) and multivariable logistic regression. RESULTS: Of the 398 included patients, 181 (45%) had malnutrition, with 34 (9%) scored as severely malnourished. The SI was significantly lower in malnourished patients than in well-nourished patients (64 ± 27 vs 72 ± 25; P = 0.002), and reductions in SI corresponded to increased malnutrition severity (P = 0.001). As a screening tool, the SI was an indicator of malnutrition risk (AUC 0.61) and performed slightly better than the more complex modified-NUTRIC score (AUC = 0.57). SI cutoffs of 101 and 43 had >90% sensitivity and >90% specificity, respectively, for the prediction of malnutrition. Patients with a low SI (≤43) had a significantly higher risk of mortality (HR = 2.61, 95% CI 1.06-6.48, P = 0.038). CONCLUSION: The frequency of malnutrition was high in this critically ill population, and it was associated with a poor prognosis. The SI could be used to assess nutrition risk in ICU patients.