Abstract

Background: Skeletal muscle index (SMI, kg/m2) derived from bioelectrical impedance analysis (BIA) is widely used to assess muscle mass. However, SMI is influenced by extracellular fluid, potentially leading to inaccurate estimation in patients with fluid imbalance. Intracellular water normalized by height squared (ICW/height2, L/m2) may reflect intracellular body cell volume and provide complementary information to BIA-derived SMI, particularly in patients with altered hydration or low nutritional reserve. Based on these considerations, we hypothesised that ICW/height2 would demonstrate stronger correlations with BMI than SMI, that the SMI–ICW/height2 correlation would differ significantly by sex, and that hydration status (ECW/TBW) would modulate this relationship. Objective: To compare ICW/height2 with SMI and evaluate their relationships with nutritional status and hydration in hospitalized patients. Specifically, we tested the hypotheses that (1) ICW/height2 correlates more strongly with BMI than SMI; (2) the SMI–ICW/height2 correlation differs significantly between sexes; and (3) ECW/TBW and sex interact to modulate the SMI–ICW/height2 relationship. Methods: This single-centre, retrospective, exploratory cross-sectional study included 477 hospitalized patients (232 men and 245 women) who underwent body composition assessment using BIA. SMI and ICW/height2 were examined in relation to body mass index (BMI) and hydration status assessed by extracellular water to total body water ratio (ECW/TBW). Correlation analyses were performed overall and stratified by sex, BMI categories (<18.5, 18.5–24.9, ≥25 kg/m2), and ECW/TBW groups (<0.36, 0.36–0.40, >0.40). Results: ICW/height2 showed a stronger correlation with BMI than SMI in both men (R2 = 0.492 vs. 0.338) and women (R2 = 0.354 vs. 0.121). A strong correlation between SMI and ICW/height2 was observed in men (R2 = 0.634), whereas the correlation was weaker in women (R2 = 0.316). In BMI-stratified analyses, the correlation between SMI and ICW/height2 increased with higher BMI in men (R2 = 0.192, 0.467, 0.791), while in women it was markedly attenuated in the low BMI group (R2 = 0.134). Stratification by ECW/TBW showed that in men, the correlation remained strong under normal hydration conditions but was attenuated in the high ECW/TBW group. In contrast, in women, the correlation was weak even within the normal ECW/TBW range (R2 = 0.336) and particularly poor in the low ECW/TBW group (R2 = 0.060). Among women with normal ECW/TBW, those with low BMI demonstrated clear discordance between SMI and ICW/height2. Exploratory four-quadrant classification using sex-specific SMI and ICW/height2 thresholds further demonstrated that patients with discordant SMI and ICW/height2 values had distinct BMI profiles, and BMI differed significantly among the four groups in both men and women (Kruskal–Wallis test, p < 0.001). Conclusions: These exploratory findings suggest that ICW/height2 and BIA-derived SMI may show systematic discordance under specific clinical conditions, particularly in underweight or fluid-imbalanced hospitalised patients. Prospective validation against reference-standard body-composition measures is required before any clinical application.

IPC Classification

Keywords