BACKGROUND:Low muscularity and malnutrition at intensive care unit (ICU) admission have been associated with negative clinical outcomes. There are limited data available evaluating the validity of bedside techniques to measure muscle mass in critically ill adults. We aimed to compare bedside methods for muscle mass assessment [bioimpedance spectroscopy (BIS), arm anthropometry and subjective physical assessment] against reference technology [computed tomography (CT)] at ICU admission. METHODS:Adults who had CT scanning at the third lumbar area <72 h after ICU admission were prospectively recruited. Bedside methods were performed within 48 h of the CT scan. Pearson's correlation compared CT muscle area with BIS-derived fat-free mass (FFM) (kg) and FFM-Chamney (kg) (adjusted for overhydration), mid-upper arm circumference (cm) and mid-arm muscle circumference (cm). Depleted muscle stores were determined using published thresholds for each method. Cohen's kappa (κ) was used to evaluate the agreement between bedside and CT assessment of muscularity status (normal or low). RESULTS:Fifty participants were enrolled. There were strong correlations between CT muscle area and FFM values and mid-arm muscle circumference (P < 0.001). Using FFM-Chamney, all six (100%) participants with low CT muscle area were detected (κ = 0.723). FFM-BIS, arm anthropometry and subjective physical assessment methods detected 28%-38% of participants with low CT muscle area. CONCLUSIONS:BIS-derived FFM using an adjustment algorithm for overhydration was correlated with CT muscle area and had good agreement with muscularity status assessed by CT image analysis. Arm anthropometry and subjective physical assessment techniques were not able to reliably detect participants with low CT muscle area.