The effect of protein administration during critical illness depends on body composition: A secondary analysis of a prospective, observational study

Background & aims: The most adequate amount of protein that should be administered to critically ill patients is still debated and diverging findings are recently accumulating. We hypothesized that the effect of protein administration might depend on the amount of muscle mass. Methods: A secondary analysis of a single-centre prospective observational study of body composition in critically ill patients. Mechanically-ventilated subjects with an expected intensive care unit (ICU) stay >72 h were enrolled. Within 24 h from ICU admission, bioimpedance-derived muscle mass (BIA MM) and rectus femoris cross-sectional area (RF CSA) were measured. The amount of proteins and calories administered on the 7th ICU day was recorded. Results: We enrolled 94 subjects (65 males, actual body weight 72.9 ± 14.4 Kg, BMI 26.0 ± 4.8 kg/m2). Actual body weight was only weakly related to BIA MM (R = 0.478, p < 0.001) and not related to RF CSA (R = 0.114, p = 0.276). A higher protein intake was associated with a reduced mortality in the highest quartile of BIA MM (OR 0.68 [0.46; 0.99] per each 10 g of proteins administered) and in the third (OR 0.74 [0.57; 0.98]) and highest quartile of RF CSA (OR 0.68 [0.48; 0.96]). Conclusion: A higher protein intake was associated with lower ICU mortality only in patients admitted with a higher muscle mass, as either assessed by BIA or muscle ultrasound.