Baseline phase angle is associated with morphofunctional status, length of stay, and hospitalization costs in severe anorexia nervosa: the EDIGRA study.

backgroundPhase angle (PhA), derived from bioelectrical impedance analysis, reflects bioelectrical properties related to cellular mass and fluid distribution and has been proposed as a marker of malnutrition severity. However, its exploratory association with morphofunctional status and hospitalization-related outcomes in severe anorexia nervosa (AN) remains insufficiently characterized. This study aimed to explore whether baseline PhA is associated with morphofunctional status, length of stay, and inpatient costs in a specialized eating disorders unit.

methodsIn this prospective cohort study, 42 female inpatients with severe AN or other specified feeding and eating disorder were assessed at admission. Patients were stratified into tertiles according to PhA. Anthropometry, body composition by bioelectrical impedance vector analysis (BIVA), intracellular and extracellular water distribution, handgrip strength, muscle and abdominal ultrasound parameters, biochemical markers, and length of stay were recorded. Hospitalization costs were estimated using standardized diagnosis-related group daily expenditure.

resultsMean PhA values were 4.0°, 4.7°, and 5.5° in the low, mid, and high tertiles, respectively. Importantly, total body weight did not differ significantly across tertiles. In contrast, body cell mass index increased progressively (5.2, 6.1, and 6.9 kg/m²). Higher PhA was associated with greater rectus femoris cross-sectional area (2.3 vs. 3.7 cm²) and higher handgrip strength (20.3 vs. 24.9 kg), consistent with more favorable muscle structure and function in unadjusted comparisons. Hydration profiles also differed: extracellular water proportion decreased (52.7% to 40.3%), while intracellular water increased (44.1% to 53.3%) across tertiles. Median length of stay declined from 58.4 to 41.3 days, with corresponding reductions in estimated hospitalization costs (€36,523 to €25,829), which should be interpreted descriptively because cost estimates were largely driven by hospitalization duration. ROC analysis showed modest discriminatory performance for prolonged hospitalization (AUC = 0.65; exploratory threshold: 4.5-4.6°).

conclusionsBaseline PhA was associated with differences in morphofunctional profiles and hospitalization trajectories despite similar body weight, suggesting that it may capture morphofunctional variability not reflected by anthropometry alone. Its associations with cellular mass, hydration distribution, muscle function, and length of stay suggest that PhA may provide complementary descriptive information within a broader clinical assessment. However, given the modest discriminative performance, sample size, and lack of multivariable adjustment, PhA should not be considered a standalone prognostic tool. Larger multicenter studies are needed to validate thresholds and formally test whether PhA provides incremental information beyond conventional anthropometric, clinical, psychological, and organizational factors. ETHICS APPROVAL: Approved by the Provincial Research Ethics Committee of Granada (SICEIA-2024-003069).